e-Learning, e-Education, and Online Training: 6th EAI International Conference, eLEOT 2020, Changsha, China, June 20-21, 2020, Proceedings, Part II [1st ed.] 9783030639549, 9783030639556

This 2-volume set constitutes the proceedings of the 6th International Conference on e-Learning, e-Education, and Online

345 104 18MB

English Pages XV, 350 [362] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

e-Learning, e-Education, and Online Training: 6th EAI International Conference, eLEOT 2020, Changsha, China, June 20-21, 2020, Proceedings, Part II [1st ed.]
 9783030639549, 9783030639556

Table of contents :
Front Matter ....Pages i-xv
Front Matter ....Pages 1-1
Application of BIM in the Course of Subgrade Construction Technology (Man-li Tian, Jie Wang, Ai-jun Jiang, Wei-wei Zhu)....Pages 3-17
On the Evolution of Knowledge Graph Abroad and Its Application in Intelligent Education (Yuliu Zhang, Bo Zhao)....Pages 18-24
Effectiveness Evaluation of Network Security Knowledge Training Based on Machine Learning (Quan-wei Sheng)....Pages 25-37
Based on the Big Data Program Language Learning Website Generation System (Quanwei Sheng)....Pages 38-53
Evaluation Model of Aquaculture Robot Technology Research Project Based on Machine Learning (Peng Cheng)....Pages 54-64
Research on the Application of Artificial Intelligence Technology in the Quality Evaluation of Experiential Physical Education (Tian Xiong)....Pages 65-76
Encryption and Compression Storage Method of Educational Resources Based on Complex Network and Big Data Analysis (Ji-shun Ma, Bin Zhao, Hao Zhang)....Pages 77-86
High-Quality Extraction Method of Education Resources Based on Block Chain Trusted Big Data (Hao Zhang, Bin Zhao, Ji-shun Ma)....Pages 87-96
Application of Open-Source Software in Knowledge Graph Construction (Qianqian Cao, Bo Zhao)....Pages 97-102
Design of Network Security Defense Knowledge Training Management Platform Under Cloud Media (Ji-yin Zhou, Chun-rong Zhou)....Pages 103-114
Design of Computer-Aided Course Teaching Control System Based on Supervised Learning Algorithm (Chun-rong Zhou, Ji-yin Zhou)....Pages 115-126
Research on the Finite Element Analysis of the Sealing Property of the Piston Used in Automobile Teaching (Hai-liang Liu, Da-wei Ding, Xin Huan, Fang-yan Yang)....Pages 127-139
Research on Multi-agent Robot Behavior Learning Based on Fuzzy Neural Network (Jun-ru Wang)....Pages 140-151
Game Model of Regional Education and Economic Development Based on Association Rule Mining Algorithm (Shan Wang)....Pages 152-164
Automatic Detection of Image Features in Basketball Shooting Teaching Based on Artificial Intelligence (Sha Yu, Jing Liu)....Pages 165-175
Design of Higher Education Aided Teaching System Based on Analysis of Economic Coupling Coordination Degree (Shan Wang)....Pages 176-188
Research on the Optimization of the Allocation of Educational Resources in Modern History Based on Demand Features (Hong-Gang Wang, Jing Xia)....Pages 189-199
Design of Learning Feedback System of Sports Training Based on Big Data Analysis (Lei Chen, Fei Gao, Gewei Zhuang)....Pages 200-214
Design of Electrical Remote Control Teaching System Based on Intelligent Ubiquitous Learning Network Model (Zhuang Gewei)....Pages 215-226
Research on the Integrated Mode of Ideological and Political Education in Colleges and Universities Based on Multivariate Data Analysis (Zhu-zhu Li, Ming-jun Cen)....Pages 227-238
Simulation Training Remote Control System of Industrial Robot Based on Deep Learning (Dan Zhao, Ming Fei Qu)....Pages 239-251
Knowledge Training System of Urban Pest Control Based on Big Data Analysis (Xin Zhang, Ming-fei Qu)....Pages 252-264
Research on Educational Data Mining Based on Big Data (Xuping He, Wensheng Tang, Jia Liu, Bo Yang, Shengchun Wang)....Pages 265-278
Research on DINA Model in Online Education (Jia Liu, Wensheng Tang, Xuping He, Bo Yang, Shengchun Wang)....Pages 279-291
Design of Data-Driven Visualization Teaching System for Preschool Basketball Courses (Y.-f. Qin, L.-z. Mo, Can Chen)....Pages 292-302
Design of Distance Multimedia Physical Education Teaching Platform Based on Artificial Intelligence Technology (Lu-zhen Mo, Yun-fei Qin, Zhu-zhu Li)....Pages 303-314
Research on 3D Printing and Its Application in CAD Teaching (Rui-can Hao, Hua-gang Liu)....Pages 315-322
Research on the Application of Blockchain Technology in Intangible Cultural Heritage Creative Design and Teaching (Liu Jun, Zhu Tiejun)....Pages 323-334
Semantic Retrieval Method of UK Education Resource Metadata in Hierarchical Cloud P2P Network (Xia Wang, Ming-Jun Li)....Pages 335-347
Back Matter ....Pages 349-350

Citation preview

Shuai Liu Guanglu Sun Weina Fu (Eds.)

340

e-Learning, e-Education, and Online Training 6th EAI International Conference, eLEOT 2020 Changsha, China, June 20–21, 2020 Proceedings, Part II

Part 2

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering Editorial Board Members Ozgur Akan Middle East Technical University, Ankara, Turkey Paolo Bellavista University of Bologna, Bologna, Italy Jiannong Cao Hong Kong Polytechnic University, Hong Kong, China Geoffrey Coulson Lancaster University, Lancaster, UK Falko Dressler University of Erlangen, Erlangen, Germany Domenico Ferrari Università Cattolica Piacenza, Piacenza, Italy Mario Gerla UCLA, Los Angeles, USA Hisashi Kobayashi Princeton University, Princeton, USA Sergio Palazzo University of Catania, Catania, Italy Sartaj Sahni University of Florida, Gainesville, USA Xuemin (Sherman) Shen University of Waterloo, Waterloo, Canada Mircea Stan University of Virginia, Charlottesville, USA Xiaohua Jia City University of Hong Kong, Kowloon, Hong Kong Albert Y. Zomaya University of Sydney, Sydney, Australia

340

More information about this series at http://www.springer.com/series/8197

Shuai Liu Guanglu Sun Weina Fu (Eds.) •



e-Learning, e-Education, and Online Training 6th EAI International Conference, eLEOT 2020 Changsha, China, June 20–21, 2020 Proceedings, Part II

123

Editors Shuai Liu Human Normal University Changsha, China

Guanglu Sun Harbin University Harbin, China

Weina Fu Hunan Normal University Changsha, China

ISSN 1867-8211 ISSN 1867-822X (electronic) Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering ISBN 978-3-030-63954-9 ISBN 978-3-030-63955-6 (eBook) https://doi.org/10.1007/978-3-030-63955-6 © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

We are delighted to introduce the proceedings of the 6th edition of the 2020 European Alliance for Innovation (EAI) International Conference on e-Learning e-Education and Online Training (EAI eLEOT 2020). This conference has brought together researchers, developers, and practitioners from around the world who are leveraging and developing information technology for educational modernization, such as artificial intelligence and big data. The theme of eLEOT 2020 was “Education with New Generation Information Technology.” The technical program of eLEOT 2020 consisted of 62 full papers, including 2 invited papers in oral presentation sessions at the main conference tracks. The conference tracks were: Track 1 – Education based Information Technology; and Track 2 – New Generation Information Technology in Education. Aside from the high-quality technical paper presentations, the technical program also featured two keynote speeches. The two keynote speakers were Prof. Yu-dong Zhang from the School of Informatics, University of Leicester, UK, who was the fellow of IET (FIET), the senior member of IEEE and ACM, and the 2019 recipient of “Highly Cited Researcher” by Web of Science; as well as Prof. Gautam Srivastava from the Department of Mathematics and Computer Science, Brandon University, Canada, who was the senior member of IEEE and popularly known in the field of data mining and big data, with more than 60 high-quality publications. Coordination with the steering chair, Prof. Imrich Chlamtac, Bruno Kessler Professor, University of Trento, Italy, was essential for the success of the conference. We sincerely appreciate his constant support and guidance. It was also a great pleasure to work with such an excellent Organizing Committee team and we thank them for their hard work in organizing and supporting the conference. In particular, the Technical Program Committee (TPC), led by our TPC co-chairs, Dr. Fei Lang from Harbin University of Science and Technology, China, and Prof. Lei Chen from Georgia Southern University, USA, who completed the peer-review process of technical papers and made a high-quality technical program. We are also grateful to the conference manager, Barbora Cintava, for her support and all the authors who submitted their papers to the eLEOT 2020 conference. We strongly believe that the eLEOT conference provides a good forum for all researchers, developers, and practitioners to discuss all science and technology aspects that are relevant to new technologies in education. In particular, all of us believe it is the right time to introduce new technologies for online education since COVID-19 is attacking the world. We also expect that future eLEOT conferences will be as successful and stimulating as indicated by the contributions presented in this volume. October 2020

Shuai Liu Guanglu Sun Weina Fu

Organization

Steering Committee Imrich Chlamtac

University of Trento, Italy

Organizing Committee General Chair Guanglu Sun

Harbin University of Science and Technology, China

TPC Chair and Co-Chairs Shuai Liu Fei Lang Lei Chen

Hunan Normal University, China Harbin University of Science and Technology, China Georgia Southern University, USA

Sponsorship and Exhibit Chair Jiazhong Xu

Harbin University of Science and Technology, China

Local Chair Wensheng Tang

Hunan Normal University, China

Workshops Chair Suxia Zhu

Harbin University of Science and Technology, China

Publicity and Social Media Chair Ao Li

Harbin University of Science and Technology, China

Publications Chairs Zhaojun Li Huiyu Zhou

Western New England University, USA University of Leicester, UK

Web Chairs Jing Qiu Weina Fu

Harbin University of Science and Technology, China Hunan Normal University, China

Posters and PhD Track Chair Rafi Judgal

Harbin University of Science and Technology, China

viii

Organization

Panels Chair Xiaochun Cheng

Middlesex University, UK

Technical Program Committee Dan Sui Jianfeng Cui Arun Kumar Sangaiah Lei Ma Wei Wei Shuai Yang Ruolin Zhou Yongjun Qin Gihong Min Xuanyue Tong Jinming Wen Yanning Zhang Khan Muhammad Lili Liang Carlo Cattani Xinyu Liu Zheng Pan Shuai Liu Shuai Wang Jiesheng He Jing Qiu Yun Lin Chengyan Li Weina Fu Jian Luo

California State Polytechnic University, Pomona, USA Xiamen University of Technology, China Vellore Institute of Technology, India Beijing Polytechnic University, China The University of Texas at Dallas, USA Changchun University of Technology, China Western New England University, USA Guilin Normal College, China Pai Chai University, South Korea Nanyang Institute of Technology, China Centre national de la recherche scientifique, France Beijing Polytechnic University, China Sejong University, South Korea Harbin University of Science and Technology, China Tuscia University, Italy Hunan Normal University, China Horizon Robotics Co. Ltd, China Hunan Normal University, China Hunan Normal University, China Hunan Normal University, China Harbin University of Science and Technology, China Harbin Engineering University, China Harbin University of Science and Technology, China Hunan Normal University, China Hunan Normal University, China

Contents – Part II

New Generation Information Technology in Education Application of BIM in the Course of Subgrade Construction Technology . . . . Man-li Tian, Jie Wang, Ai-jun Jiang, and Wei-wei Zhu

3

On the Evolution of Knowledge Graph Abroad and Its Application in Intelligent Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yuliu Zhang and Bo Zhao

18

Effectiveness Evaluation of Network Security Knowledge Training Based on Machine Learning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quan-wei Sheng

25

Based on the Big Data Program Language Learning Website Generation System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quanwei Sheng

38

Evaluation Model of Aquaculture Robot Technology Research Project Based on Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peng Cheng

54

Research on the Application of Artificial Intelligence Technology in the Quality Evaluation of Experiential Physical Education . . . . . . . . . . . . Tian Xiong

65

Encryption and Compression Storage Method of Educational Resources Based on Complex Network and Big Data Analysis . . . . . . . . . . . . . . . . . . Ji-shun Ma, Bin Zhao, and Hao Zhang

77

High-Quality Extraction Method of Education Resources Based on Block Chain Trusted Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hao Zhang, Bin Zhao, and Ji-shun Ma

87

Application of Open-Source Software in Knowledge Graph Construction . . . . Qianqian Cao and Bo Zhao

97

Design of Network Security Defense Knowledge Training Management Platform Under Cloud Media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ji-yin Zhou and Chun-rong Zhou

103

Design of Computer-Aided Course Teaching Control System Based on Supervised Learning Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chun-rong Zhou and Ji-yin Zhou

115

x

Contents – Part II

Research on the Finite Element Analysis of the Sealing Property of the Piston Used in Automobile Teaching . . . . . . . . . . . . . . . . . . . . . . . . Hai-liang Liu, Da-wei Ding, Xin Huan, and Fang-yan Yang

127

Research on Multi-agent Robot Behavior Learning Based on Fuzzy Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jun-ru Wang

140

Game Model of Regional Education and Economic Development Based on Association Rule Mining Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . Shan Wang

152

Automatic Detection of Image Features in Basketball Shooting Teaching Based on Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sha Yu and Jing Liu

165

Design of Higher Education Aided Teaching System Based on Analysis of Economic Coupling Coordination Degree. . . . . . . . . . . . . . . . . . . . . . . . Shan Wang

176

Research on the Optimization of the Allocation of Educational Resources in Modern History Based on Demand Features . . . . . . . . . . . . . . . . . . . . . . Hong-Gang Wang and Jing Xia

189

Design of Learning Feedback System of Sports Training Based on Big Data Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lei Chen, Fei Gao, and Gewei Zhuang

200

Design of Electrical Remote Control Teaching System Based on Intelligent Ubiquitous Learning Network Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhuang Gewei

215

Research on the Integrated Mode of Ideological and Political Education in Colleges and Universities Based on Multivariate Data Analysis. . . . . . . . . Zhu-zhu Li and Ming-jun Cen

227

Simulation Training Remote Control System of Industrial Robot Based on Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dan Zhao and Ming Fei Qu

239

Knowledge Training System of Urban Pest Control Based on Big Data Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xin Zhang and Ming-fei Qu

252

Research on Educational Data Mining Based on Big Data . . . . . . . . . . . . . . Xuping He, Wensheng Tang, Jia Liu, Bo Yang, and Shengchun Wang

265

Contents – Part II

Research on DINA Model in Online Education. . . . . . . . . . . . . . . . . . . . . . Jia Liu, Wensheng Tang, Xuping He, Bo Yang, and Shengchun Wang

xi

279

Design of Data-Driven Visualization Teaching System for Preschool Basketball Courses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Y.-f. Qin, L.-z. Mo, and Can Chen

292

Design of Distance Multimedia Physical Education Teaching Platform Based on Artificial Intelligence Technology . . . . . . . . . . . . . . . . . . . . . . . . Lu-zhen Mo, Yun-fei Qin, and Zhu-zhu Li

303

Research on 3D Printing and Its Application in CAD Teaching . . . . . . . . . . Rui-can Hao and Hua-gang Liu

315

Research on the Application of Blockchain Technology in Intangible Cultural Heritage Creative Design and Teaching . . . . . . . . . . . . . . . . . . . . . Liu Jun and Zhu Tiejun

323

Semantic Retrieval Method of UK Education Resource Metadata in Hierarchical Cloud P2P Network. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xia Wang and Ming-Jun Li

335

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

349

Contents – Part I

Education Research with Information Technology Empirical Research on the Creative Design Talents Cultivation of Sino-Foreign Cooperative Education Project in Local Engineering Universities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Li Yuan and Zhu Tiejun

3

Evaluation Model of Case Teaching Effect of Engineering Cost Based on Data Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xiao Qi-rong

16

Data Analysis of Cost Engineer Qualification Examination System Based on Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xiao Qi-rong

29

Performance Evaluation of Integrated Circuit Industry Talent Training Based on BP Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xia Jing and Wang Hong-Gang

43

The Satisfaction Evaluation Model of Course Resources of Automobile Maintenance Major Based on Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Lu and Y. Zhou

55

Evaluation Method of Multimedia Art Teaching Courseware Playback Effect Based on Data Envelopment Analysis . . . . . . . . . . . . . . . . . . . . . . . Yong-ji Zhou and Shou-qing Lu

67

Design of Teaching Platform for Visual Programming of Industrial Robot Based on PBL and Multimedia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peng Cheng

76

Construction of Network Course System of Construction Machinery Specialty Based on Cloud Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xu You-jun

88

Research on Innovation and Entrepreneurship Education Model of Higher Vocational Colleges Based on Internet Perspective . . . . . . . . . . . . . . . . . . . You-jun Xu

99

Effect Analysis of Physical Education Course Based on Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tian Xiong

113

xiv

Contents – Part I

Construction of Dual System Teaching System for Automobile Detection and Maintenance Under 1+X Certificate System . . . . . . . . . . . . . . . . . . . . . Hai-jing Zhou and Shao-feng Han The Construction of the Remote Interactive Platform of the Practical Training Teaching in the Employment Domain of Colleges and Universities Under the 1 + X Certificate System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hai-jing Zhou and Shou-liang He

124

137

Education Reform of Construction Specialty in the Context of Upgrading of Intelligent Construction Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wei-wei Zhu and Man-li Tian

148

Design of Traditional Teaching Method of Micro-teaching Based on Blended Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fang-yan Yang

159

Personalized Recommendation Technology of Network Teaching Resources Based on Ant Colony Algorithm . . . . . . . . . . . . . . . . . . . . . . . . Hai-long Liu and Lei-lei Jiang

171

Dynamic Adjustment Mechanism of Intelligent Classroom Learning Resources in Universities Based on Network Teaching Platform . . . . . . . . . . Lei-lei Jiang, Rong Xu, and Hai-long Liu

181

Design of APP Learning Platform for Oil Storage Tank Mechanical Cleaning Technology Course Based on Mobile Terminal . . . . . . . . . . . . . . . Jun-ru Wang

193

Innovation and Entrepreneurship Education System of New Engineering Talents Based on Knowledge Base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yan Wang, Fang Sonh, and Hai-hong Bian

204

Design of Embedded Course Teaching System Based on Cognitive Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hai-hong Bian, Yi-qun Zhong, and Yan Wang

217

Research on Online Physical Education Micro Course System Based on Improved Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chang-min Lv, Xue-ping Zhang, and Jun-peng Ji

230

Design of Collaborative Teaching Mode of Online and Offline Based on Supervised Learning Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xue-ping Zhang, Fei-fei Qin, Ting Zhang, and Chang-min Lv

242

Multi Module Integration Method of Students’ Habitual Learning Mode Based on Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhan-yong Chen and Jun Guo

254

Contents – Part I

Curriculum Quality System Model of Entrepreneurship and Innovation Education in Vocational Colleges Across the Straits Based on Internet+. . . . . Jun Guo and Zhan-yong Chen Design of General Integrated Teaching System for Operational Research . . . . Jing Liu, Jun-feng Qiao, and Sha Yu

xv

271 286

Analysis of the Conflict Model of Education and Culture in China and Britain Based on Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xia Wang

298

The Information Grading Management System of College Students Based on Deep Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tao Jiang and Ye Song

308

Teaching Quality Evaluation Method Based on Multilayer Feedforward Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ye Song and Tao Jiang

318

Research on Planning Methods of Students’ Professional Development Trajectory Based on Big Data Forecast . . . . . . . . . . . . . . . . . . . . . . . . . . . Ming-jun Cen and Zhu-zhu Li

329

Interactive Design and Application of Preschool Education IH5 Web Page Based on Cloud and 2–3.5-Year-Old Children’s Psychology. . . . . . . . . . . . . Zhu Tiejun, Ma Wenqing, and Zheng Yue

341

Research on the Dimensions of Art Design Education in Taiwan Shih Chien University . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhu Tiejun and Zhang Linlin

351

Study on Teaching of Engineering Design Course with 3D Modeling Software and 3D Printer in International Training Course. . . . . . . . . . . . . . . Rui-can Hao, Hua-gang Liu, and Shang Wang

362

The Integrated Design of ‘Industry-University-ResearchApplication-Cultivation’ Based on the Course of ‘Mechanical Design Foundation’ in Higher Vocational Colleges . . . . . . . . . . . . . . . . . . . . . . . . Qing-song Zhu and Rui-can Hao An Investigation of Intercultural Communicative Competence Among Master’s Graduate Students of Non-English Major in the Context of Content-Based English Instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Qingling Wang and Qingying Zhou Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

370

376

387

New Generation Information Technology in Education

Application of BIM in the Course of Subgrade Construction Technology Man-li Tian(B) , Jie Wang, Ai-jun Jiang, and Wei-wei Zhu School of Road Bridge and Architecture, Chongqing Vocational College of Transportation, Chongqing 402247, China [email protected]

Abstract. The subgrade construction technology course is a core course of the professional engineering of bridges and bridges in higher vocational colleges. Due to the high professional technical requirements, strong professional practice, wide coverage, and rich professional knowledge content, it has led to the teaching of existing classroom teaching methods. The increased difficulty is not conducive to students’ good understanding of three-dimensional structures and complex construction techniques, and it is difficult to combine teaching and learning closely. In order to better carry out the teaching of subgrade construction technology courses, combined with the application of BIM technology in the teaching of subgrade construction technology courses, according to the actual conditions of school teaching, using BIM technology and guiding students to create various bridge models, innovative teaching methods and methods, Improved students’ interest in learning and enhanced their practical skills. Keywords: BIM · Subgrade construction technology · Course teaching · Teaching model

1 Introduction In the field of road construction, the application of BIM technology is becoming increasingly mature in foreign countries, and it has also set off a wave of BIM technology application in China. However, how to combine BIM technology and concepts with the teaching of roadbed construction technology courses, how to improve the teaching effect and efficiency of roadbed construction technology courses, and promote the transformation of the education world are far-reaching topics worthy of joint discussion and research by the education and road industry [1]. Taking the teaching reform of subgrade construction technology course as an example, the feasibility of applying BIM technology and its superiority compared with traditional teaching methods are explored. The subgrade construction technology course is a core and backbone professional course for vocational students majoring in road and bridge engineering technology. It mainly studies the construction methods, construction technology, and construction organization and management of bridges. The subgrade construction technology is complex © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 3–17, 2020. https://doi.org/10.1007/978-3-030-63955-6_1

4

M. Tian et al.

and updated quickly, the environment is difficult, the engineering volume is large, and the content is important, which places high requirements on the teaching of subgrade construction technology courses. The subgrade construction course of our college has gone through the teaching method reform of “construction drawing as carrier”, and the teaching reform and practice of subgrade construction technology course with the goal of establishing network course has accumulated rich teaching materials and resources, with a strong foundation [2]. On this basis, in order to further solve the problems of the implementation of construction drawing teaching in the course of bridge construction and the effect and efficiency in the process of construction drawing teaching, BIM Technology is used to further deepen the curriculum reform. Since 2013, universities such as Tsinghua University, Huazhong University of Science and Technology, Shenyang Jianzhu University, and Sichuan University have established BIM-related institutions. The Shaanxi Railway Engineering Vocational and Technical College also established the BIM Technology Application Center in 2013. It can be seen that universities have also attached importance to the research and application of BIM technology. With the establishment of BIM technology research institutions, the strengthening of talent teams, and the application of engineering projects, it will surely guide the teaching reform in the process of talent training.

2 Design of Teaching Methods for Subgrade Construction Technology Course 2.1 Making the Teaching Implementation Plan of Subgrade Construction Technology Course Based on a complete set of construction drawings under construction, a 1:1 threedimensional entity model is built according to the actual size of the construction drawings by using Revit and other software. The model can not only combine the whole typical project together, but also can be divided into parts, enriched the teaching materials, enriched the teaching means, and achieved the synchronous update of the teaching content and the specific construction technology and construction process on the construction site. Integrate the course structure of subgrade construction technology, integrate the original foundation engineering course and subgrade construction technology course into one course. Extend the class time of the original subgrade construction technology course, and add the content of bridge foundation construction in class time based on the original teaching content [3]. Following the natural sequence of subgrade construction, we first start with the construction of the road foundation and gradually transition to the pavement abutment. Finally, we talk about the content of the upper main beam and the pavement. This improves the logic of the subgrade construction technology course from the perspective of construction. And good engineering connection issues are conducive to improving students’ logical training of project construction. Plan the training of nuclear occupation ability based on BIM technology for subgrade construction technology course construction drawing reading. As one of the main tasks of the subgrade construction technology course is to train students’ ability to read

Application of BIM

5

construction drawings, only after reading the construction drawings can they understand the structure of the construction drawings accurately, and then they can guide the site construction according to the construction drawings and accurately convert the drawings and structures Engineering in kind. Combined with the inspiration of German action-oriented teaching method, according to the actual and objective conditions of our institute, in the implementation of the bridge construction teaching reform organization, our overall plan is based on the construction drawing reading ability and construction ability training ideas: First, members of the subgrade construction technology course reform team are familiar with a set of typical subgrade drawings, follow the learning rules from simple to complex, and from single to holistic, and use the development and training project of the complete construction drawing in use. According to the actual size of the drawing, a three-dimensional solid model of 1:1 is built using revit. When modeling, there are not only the overall structural effect diagram, but also the detailed steel bar layout diagrams of various parts. Second, based on the familiar drawings and the ability to establish subgrade model, train students to learn to use Revit and other modeling software, so that students can skillfully view the established model and drawings at the same time, and carry out comparative analysis, so that students gradually have the ability to accurately create three-dimensional BIM drawing according to the construction drawing, and can achieve in the training process Goal of proficiency [4]. The calculation of reinforcement beam quantity in subgrade stage is shown in Fig. 1. Third, according to the needs of specific teaching situations, after each big situation is completed. Select a construction detail of a typical drawing to guide students to use Revit to accurately build the solid model of each component of the bridge according to the size. Because in the process of model making, it is necessary to strictly refer to the construction drawings, accurately model, and operate the computer room. On the one hand, it improves the students’ ability to solve problems by hands. On the other hand, the model established is visible, and it is easy to check whether the students can really accurately understand the construction drawings. In the course of subgrade construction technology, some teaching situations, such as pile cap, pier, abutment and main beam, are repeated in the teaching practice, so as to improve the students’ reading ability of construction drawings [5]. Strengthen the construction of equipment for the training conditions of roadmap construction technology courses, purchase high-end computers with independent graphics cards that meet the needs of 3D modeling, and establish professional computer rooms and 3D printers equipped with 3D modeling software. Students who have a good level of modeling can allow them to print their work and affix their own information cards as a reward. A circle of cabinets is provided around the computer room. The cabinets store eleven sets of construction drawings prepared for our learning environment. Each set of drawings is 50 copies. It is equipped with various design and construction specifications for roadbeds, construction manuals, and 10 copies of each specification or manual. Equipped with a calculator, drawing tools, and cropping gadgets of 50 copies each for easy reference.

6

M. Tian et al.

Structur al joint

(a)Cantilever type

High strength composite

Supporting joist

(b)Joist type Fig. 1. Calculation of reinforcement beam quantity in subgrade stage

2.2 Establishing a Teaching Model for Subgrade Construction Technology Courses The basic structure of the teaching model of subgrade construction technology course based on BIM Technology is mainly a whole teaching model of subgrade construction technology course, which is composed of students as the main body, students’ learning activities in the classroom as a small structural unit, with the basic theory of independent learning proposed by social cognitive school and Zimmerman as the axis.

Application of BIM

7

The teaching model is mainly composed of three levels, that is, the basic attainment level-the knowledge promotion level-the inquiry expansion level (as shown in Fig. 3) [6]. Students’ autonomous learning in the classroom is performed in the order shown in the figure. In order to reach the level of inquiry development, students need to first reach the level of knowledge improvement, and knowledge improvement requires students to learn the various advanced rules required by the level of inquiry. For these advanced rules, students must reach the basic achievement level, that is, learn to identify and firmly grasp the specific concepts, rules and definitions (Fig. 2).

Inquiry and expansion Knowledge and ability to improve as a prerequisite

Knowing and improving It is necessary to take the basic standard as the prerequisite

Basic compliance The prerequisite of knowledge learned

Fig. 2. Structure of teaching model of subgrade construction technology course

The main part of the teaching model of subgrade construction technology course includes three closed main loops. The first one is to determine the learning objectives, stimulate learning motivation, recall the existing knowledge in the mind, self-study textbooks, basic standards, consolidate basic exercises, students’ self summary, and students’ self-evaluation. The first loop is designed for the whole class. The main content of learning is the basic concepts, laws, axioms and other declarative knowledge in the textbook. The second main loop consists of determining learning objectives, stimulating learning motivation, recalling the existing knowledge in the mind, self-study teaching materials, improving knowledge, consolidating knowledge practice, learning self summary and self-evaluation of students. This loop is mainly designed for middle-level and above students with certain foundation. The main content of learning is to learn advanced rules such as mathematical thoughts, methods and strategies on the basis of understanding the basic concepts, laws, axioms and other declarative knowledge. The third main loop is composed of determining learning goals, motivating learning, recalling existing knowledge in the mind, self-study textbooks, exploring and expanding, consolidating

8

M. Tian et al.

and expanding exercises, student self-summary, and student self-evaluation [7]. This loop is mainly for students with good academic performance, quick thinking and diligent thinking. The main content of learning is to transfer the knowledge from textbooks to real life to solve problems, so as to develop students’ innovative thinking, expand students’ horizons, The purpose of solving real life problems. Each level of learning includes three closed sub loops. The first sub loop consists of determining learning objectives, stimulating learning motivation, recalling existing knowledge in the mind, self-study textbooks, self-study inspection, standard practice consolidation, students’ self summary, students’ self-evaluation and other links. This sub loop shows that students can achieve the goal by reading and understanding the content of the textbook and completing the basic standard exercises correctly after setting the learning goal. In this smooth situation, students can independently complete the contents of reading materials, internalize basic knowledge, and complete the after-school basic standard exercises. Teachers can patrol in the classroom and do a good job in guiding and maintaining classroom discipline. The second sub loop adds collective discussion on the basis of the first sub loop, which indicates that students encounter certain obstacles when they read and understand the content of the textbook independently, but these obstacles in the process of self-study can be solved through collective discussion. Such collective discussion is mainly carried out among learning groups, so teachers only maintain classroom discipline and guide students’ learning. Self study textbooks are mainly completed by students themselves or through collective discussion within the learning group, and students master the autonomy of learning [8]. The third sub-loop adds teachers to explain this link on the basis of the second sub-loop. It shows that after self-study materials and group discussions in the study group, there is still a certain amount of knowledge understanding and internalization problems that have not been resolved. At this time, teachers need to give lectures. Teachers’ lectures help students remove learning obstacles and complete self-study tasks. Achieve self-study goals. Of course, if after the teacher’s lecture, there are still a small number of students who are unable to complete the task of self-study materials, the teacher must find out the reason why the students’ learning is blocked, or re-teach or stop teaching. The specific implementation steps of the teaching model of subgrade construction technology course are as follows: Step1: Determine teaching objectives (1) The content of the basic teaching materials and the judgment standard to achieve the learning objectives; (2) Tips on the specific learning behaviors of each link in the process of students’ learning activities; (3) Consolidate exercises. Step2: Motivate learning Students are not always passive in the process of stimulating the learning motivation. They need to psychologically imply that they want to learn the content of this lesson

Application of BIM

9

and can learn from each other. The group members in the group should encourage each other to ensure that they and their group Efficient completion of learning tasks. Step3: Let students recall what they have in mind In the process of students’ independent recall of existing knowledge, teachers can use language, gesture or other means to situationally restore the abstract text knowledge, so that students can have a sense of immersive experience in the process of knowledge extraction, which is more helpful for students to quickly form a road map into a certain “knowledge area” in their minds, and successfully reach the destination. Step4: Guide students’ self-study ability Students’ self-directed learning can’t just start with textbooks. Teachers need to remind students to adopt a combination of self-directed learning and free discussion in groups so that they can learn more knowledge in one lesson. And when the self-learning classroom teaching was first tried in the second grade, teachers needed to guide students to learn together instead of letting students learn freely, so that they could learn more systematic and solid basic knowledge. Step5: Self study examination After learning the contents of the textbook independently, students should do exercises according to the passing exercises given by the teacher or the exercises at the back of the textbook and compare them with the true answers given by the teacher, calculate the accuracy rate or the degree of understanding and application, and feed back the comparison results in time. Members of the group urge each other to check whether their learning objectives have been achieved? What is the correct rate of exercises? What are the obstacles to learning goals? What are the difficulties of the poor students in the group. In the process of checking whether the exercises are right or wrong, the teacher can invite the students with good, poor and middle grades to write their solutions on the blackboard, and check whether their solutions are correct together with the whole class. In this way, the teacher can not only understand the students’ autonomous learning in the class, but also facilitate the discussion among the whole class Mistakes in the process of doing the questions. Step6: Organize discussions The discussion among the members of the group can begin with the solution of the exercises of a certain classmate in the group, or start with the important and difficult points of study given by the teacher in this class. The discussion about this lesson between the students in the whole class can be started under the guidance of the teacher around the solution process of the practice questions shown by several students on the blackboard. How can a student comment on a student’s problem-solving process? Where is it calculated? Can also point out the reasons and crux of a student’s error proneness? And ask students with strong language organization skills and agile thinking to lead the

10

M. Tian et al.

classmates to make a summary, so that the students in the class know exactly what is learned in this lesson and where they are in this chapter Why is it important? Where is the difficulty? How to overcome it? and many more. Step7: Key lectures by teachers In the form of students’ own control of learning progress, Depends mainly on students to succeed, Not through machines and other technical means. “Group teaching complemented by the frequent feedback and individual corrective help required by each student”. It is the essence of learning. Bloom’s mastery of learning is based on a new perspective on students, he thinks “Almost all people can learn what one can learn in the world by providing proper conditions of the past and the present”. This is in line with the promotion of new curriculum reform in China “Commonness + individuality” Cultivated spirit is consistent, That is to say, in the stage of basic compliance, Require each student to meet standards “Mastery Learning”. The basic implementation procedure of teaching is: Analysis of teaching materials - reorganization of teaching materials - Design of “unit feedback - correction procedure” - final evaluation. However, in the classroom teaching situation of autonomous learning, the main task of teachers is to use the “feedback correction procedure” repeatedly and frequently for the students who fail to reach the standard, so that each student can finally complete the learning task at the basic standard stage. 2.3 Realization of Courses in Subgrade Construction Technology According to the three-level structure of teaching model, the realization of subgrade construction technology course teaching also presents three-level auxiliary structure, namely three-level auxiliary teaching cycle. The first loop is to attract students’ attention, inform students of learning goals, present stimulating materials, provide learning guidance, lead basic exercise questions, provide homework-specific feedback, evaluate homework, teacher summary, and evaluate student knowledge and skills. The second loop is to attract attention, inform students’ objectives, present stimulus materials, provide learning guidance, lead out ability improvement exercises, provide correct feedback on assignments, evaluate assignments, teachers’ summary, evaluate student processes and methods, etc. [9]. The third loop is to draw attention, inform students of goals, present stimulating materials, provide study guidance, lead inquiry and exploration exercises, provide correct feedback for assignments, evaluate assignments, teacher summary, and evaluate students’ attitudes and values. The teaching implementation process of subgrade construction technology course is shown in Fig. 3. The implementation steps of the roadbed construction technology course are as follows: Step1: Attract attention At the beginning of the class, students should be given stronger and more novel stimuli, help students to condense various other thinking activities before class, and let students’ attention be directed to the classroom quickly.

Application of BIM Provide learning guidance 1

Lead out basic exercises

11

Provide correct feedback on the job

Present stimulus material

Inform learning objectives

Attract attention

Provide learning guidance 2

Lead out basic exercises

Provide correct feedback on the job

Provide learning guidance 3

Lead out basic exercises

Provide correct feedback on the job

Go to the nex t topi c

Fig. 3. Flow chart of subgrade construction technology teaching implementation

Step2: Inform students of goals The purpose of the teacher to inform the students’ learning objectives is to establish an expectation of the behavior obtained as a result of the learning. It is the main role that learners have in anticipating learning outcomes, enabling them to match their behavior with the type of “correct” behavior they expect. Therefore, reinforcement in the form of information feedback has a further role in confirming learners’ expectations. Step3: Present stimulus material In the course of the roadbed construction technology course, advance arrangements must be made for the nature, method, and appropriate time of presenting the stimulating materials. It is important to choose the right time to present the stimulating materials. Generally, a stimulating learning material will be presented before the students learn autonomously, in order to stimulate the students’ curiosity, or present a class when students encounter difficulties in the process of autonomous learning Supplementary materials to help students overcome difficulties, or give a type of extended or supplementary learning materials after learning to deepen students’ knowledge of this type of knowledge. Step4: Provide learning guidance According to the different levels of students’ autonomous learning, teachers should give corresponding learning guidance, that is, at the level of basic standards, teachers

12

M. Tian et al.

should give “master learning” teaching guidance; at the level of knowledge and ability improvement, teachers should give “scaffolding” teaching guidance; at the level of exploration and expansion, teachers should give “insight learning” teaching guidance. Step5: Elicit a practice question Depending on the level of student’s autonomous learning, teachers need to draw different consolidation exercises. At the basic achievement level, students should present basic practice questions that consolidate their knowledge so that they can form a good knowledge structure. In the knowledge promotion level, teachers will give students some more difficult questions than the basic achievement level exercises to improve students’ hands-on ability and once again consolidate the knowledge of the basic achievement level. At the level of inquiry development, the teacher will give practice exercises with fixedness and flexibility to achieve the ability to train students to comprehensively use knowledge in the first two levels of learning. Step 6: Provide correct feedback on the job After finishing some exercises, the students will determine whether the answers of some exercises are correct or not through the discussion and comparison between groups, so the teacher must do one thing is to tell the students whether the answers of each exercise they have done are correct or not. But this does not mean that teachers must use the words “right”, “wrong”, “right” or “incorrect” [10–13]. In the classroom, teachers can show whether the exercises are correct or not through follow-up teaching, or use other subtle tips such as nodding, smiling or scanning to show whether the exercises are correct or not. Many times, the exercises in the basic standard level have been learned by students in their early life, so it is unnecessary to explain all the exercises in detail at this time. Step7: Teacher’s assessment and assignment Homework assessment is an important way for teachers and students to interact. Positive homework assessment can help students understand themselves, build confidence, and narrow the distance between teachers and students. Therefore, we need to give more encouragement to students’ homework evaluation. Make the evaluation warm and effective: We must pay attention to the development of students ‘personality and the cultivation of students’ willpower. Teachers should make full use of the link of homework evaluation, while guiding students to learn mathematical knowledge, even if it nourishes the students’ hearts, organically infiltrate learning interests, learning attitudes, and the cultivation of good learning habits; timely adjust student learning through homework evaluation, Guide them to learn self-analysis, self-appreciation, constantly improve themselves, and grow up healthily.

Application of BIM

13

Step 8: Teacher summary and evaluation At the end of a class, teachers need to make a summary of what they have learned in this class, so that students can form a logical vein of knowledge in their heads, which is not only convenient for memory storage, but also can optimize the knowledge structure of students, and easy to recall and extract when they are used in the future. The evaluation of teachers should be comprehensive, but also focus on the students’ learning process, including “knowledge and skills”, “process and method”, “emotional attitude and values”, etc. But especially for the students who encounter obstacles or difficulties in the learning process, teachers should especially explain the methods used to solve the problems, or the creative way of thinking. And the students who have successfully solved difficult problems are praised and affirmed, and all students are set an example to encourage them to learn and improve themselves.

3 Case Analysis 3.1 Pilot Arrangements According to the specific teaching arrangements, pilot projects were conducted in the subgrade construction courses of the 2013 and 2014 grades of subgrade construction technology. There are 12 classes in this major. Three classes were selected for the pilot. Three teachers led a pilot class and an ordinary class, each with about 45 students. At the end of the semester, through the comparative analysis of the final exam results, the traditional subgrade construction technology course teaching method and the BIMbased subgrade construction technology course teaching method were used to collect the experimental classes and non-experimental classes from the average score and the question type score for analysis. 3.2 Data Analysis According to the arrangement of the pilot project, the collected final score data of students are sorted out and the following results are obtained. It can be seen from Fig. 4 that the average scores of the subgrade construction technology courses in the pilot classes are 15.8 points higher than those in the non-pilot classes, which is 24.2% higher than the same period last year. In addition, from the perspective of question type, there is no significant difference between the scores of blank filling questions in pilot class and non pilot class. The average score of sample questions in pilot class is 7.9 points higher than that in non pilot class, 32.1% higher than that in the same period last year, and the average score of map reading questions in pilot class is 11.2 points higher than that in non pilot class, 46.3% higher than that in the same period last year. It is because filling in the blanks is mainly based on memorization assessment content, and the application of BIM technology is not very obvious to improve students’ memorization content, but the average score of case analysis questions and picture recognition problems is obviously improved, especially the picture recognition problems. The range reached 46.3%. This shows that the application of BIM technology

14

M. Tian et al. 90 80 70 60 50 40 30 20 10 0

Article method traditional method

Average score of case questions

Average score of map reading questions

Average

Average score of filling in blank

81.2

12

32.5

35.4

65.4

10

24.6

24.2

Teaching methods of this paper Traditional teaching methods

Fig. 4. Statistics of final exam scores of 2013 students

in the teaching process of subgrade construction technology can significantly improve students’ understanding of subgrade structure and construction process, and further enhance students’ ability to identify drawings and optimize construction plans. 90 80 70 60 50 40 30 20 10 0

Article method traditional method

Average score of case questions

Average score of map reading questions

Average

Average score of filling in blank

88.2

11

34.5

38.5

66.1

9

22.6

25.2

Teaching methods of this paper Traditional teaching methods

Fig. 5. 2014 statistics of final examination results of students

It can be seen from Fig. 5 that the average scores of the subgrade construction technology courses for the 2014 pilot classes are 22.1 points higher than the non-pilot classes, 33.4% higher than the same period last year and 11 percentage points higher than the 2013 class. The scores of fill-in-the-blank questions in the 2014 pilot class and

Application of BIM

15

non-pilot classes are not much different from those in the 2013 class; the average score of the case questions in the 2014 pilot class is 11.9 points higher than that of the non-pilot class, which is 52.7% higher than that of the previous year, and 20 higher than that of 2013 Multiple percentage points; the average score of the picture recognition questions for the 2014 pilot class was 13.3 points higher than that of the non-pilot class, which was 52.8% higher than the same period of the previous year, which was 6 percentage points higher than the 2013 class. The main task is to fill in the blank. There are many assessment contents relying on memory. The application of BIM Technology is not obvious to improve students’ memory content, but the average scores of case analysis questions and map reading questions are obviously improved, especially map reading questions. 90 80 70 60 50 40 30 20 10 0 Average score of case questions

Average score of map reading questions

Average

Average score of filling in blank

Final grade of 2013

88.2

11

34.5

38.5

Final grade of 2014

81.2

12

32.5

35.4

Final grade of 2014 Final grade of 2013

Fig. 6. Comparison of teaching methods in this article

It can be seen from Fig. 6 and Fig. 7 that compared with 2014 and 2013, the performance improvement of 2013 and 2014 non pilot classes is very limited, and the improvement of teaching effect is not obvious. However, the scores of 2013 and 2014 pilot classes increased significantly, especially the scores of map and case questions in 2014 pilot classes increased significantly. This shows that in the application of BIM technology in the teaching process of subgrade construction technology courses, due to the improvement of teachers’ teaching ability, the continuous improvement of teaching materials and teaching design, and the continuous improvement and updating of teaching equipment, students can significantly improve the students’ The ability to learn further enhances the students’ ability to identify drawings and the preparation and improvement of construction plans, improve their cognition and understanding, and improve the efficiency and effectiveness of learning.

16

M. Tian et al. 90 80 70 60 50 40 30 20 10 0 Average

Average score of filling in blank

Average score of case questions

Average score of map reading questions

Final grade of 2013

66.1

9

22.6

25.2

Final grade of 2014

65.4

10

24.6

24.2

Final grade of 2014 Final grade of 2013

Fig. 7. Comparison of traditional teaching methods

4 Concluding Remarks This paper studies the application of Bim in the course of subgrade construction technology. After several rounds of reform in the course of subgrade construction technology, certain results have been achieved. How to enrich the core of the course teaching, fundamentally improve the students ‘core competence in construction drawing reading, and enhance students’ knowledge level and ability level. The three-dimensional modeling of the whole bridge is carried out by Revit, Dassault and other software, the two-dimensional drawings are transformed into three-dimensional entities, the complex subgrade structure is three-dimensional and visualized, the process logic problems are simplified and clear, the problems of construction drawing reading in the teaching process, the inspection of construction process understanding effect and the teaching effect are difficult to grasp are solved, and the teaching effect is achieved through the pilot, However, the requirements of teaching environment and equipment are high, and the investment in teaching is relatively large. In the process of promotion, each college should consider the use in combination with the actual conditions.

5 Fund Projects Research on the Application of BIM Technology in the Course Reform of Subgrade Construction and Construction (CJJY201912).

References 1. Qin, S.J., Ouyang, J.L., Mo, W.S.: Exploration and practice of teaching reform on architectural natural lighting. Exp. Technol. Manage. 35(8), 182–186+210 (2018)

Application of BIM

17

2. Wang, Z., Jiang, Z.T.: Exploration and practice of research-oriented teaching in the course of solid state physics. Coll. Phys. 36(10), 57–60 (2017) 3. Lu, X.W., He, G.Z.: Exploration on the teachers and students interactive teaching methods in the law practice network course. Adult Educ. 37(2), 51–53 (2017) 4. Shen, W., Mai, Y.F., Qian, W.: Hydraulic transmission teaching reform based on engineering education certification. Chin. Hydraul. Pneum. 58(8), 73–78 (2017) 5. Dai, Y.: The application of BIM Technology in the project teaching of architectural courses. Kexue yu Xinxihua 52(27), 124 (2017) 6. Zhao, H., Liu, M.X., Liu, F.Y.: Study on a new daylighting technology for school buildings. Build. Sci. 34(8), 94–99 (2018) 7. Han, R., Liu, D.P., Shao, D., et al.: Research on the teaching mode of multi professional collaboration architectural design based on BIM. Res. Exp. Lab. 36(4), 174–178+266 (2017) 8. Luo, M.L.: The application of BIM in the teaching reform of college curriculum system. Course Educ. Res. 4(50), 55+108 (2017) 9. Song, S.J.: On teaching reform of architectural majors of higher vocational colleges under concept of Wisdom City. Vocat. Tech. Educ. 37(32), 38–40 (2016) 10. Jia, C.Y., Liu, L.M.: On the practice of mathematics teaching reform for construction and surveying engineering technology specialty in higher vocational colleges. Vocat. Tech. Educ. 36(32), 46–48 (2015) 11. Szwarkowski, D., Pilecka, E.: BIM technology in geotechnical engineering in terms of impact high building “Mogilska Tower” in Cracow of existing building development. Tech. Sci. 3(20), 297–309 (2017) 12. Yang, Y., Ha, M.H.: Study on the influence factors of teaching quality of BIM in the mode of promoting teaching and learning by competition. J. Hebei Univ. Eng. (Nat. Sci. Edn.) 35(3), 106–109 (2018) 13. Zhang, M., Jiang, Y.C., Meng, J.: Application of BIM technology in architectural decoration design course teaching. Heilongjiang Sci. 9(7), 60–61 (2018)

On the Evolution of Knowledge Graph Abroad and Its Application in Intelligent Education Yuliu Zhang1

and Bo Zhao1,2(B)

1 School of Information Science and Technology, Yunnan Normal University,

Kunming 650500, YN, China [email protected], [email protected] 2 Key Laboratory of Educational Information for Nationalities, Ministry of Education, Yunnan Normal University, Kunming 650500, YN, China

Abstract. The development of education is more and more dependents on the intelligent learning support services in today’s information age. The intelligent education supported by artificial intelligence (AI) has drawn more attention. In particular, knowledge graph (KG) becomes the key to promote the development and innovation of education with the developing of AI. So, the evolution of the research hotspots of KG is reviewed based on the literature of Web of Science in this study. We explore the development of intelligent education on different aspects, including educational KG, cognitive diagnosis and personalized education. Keywords: Knowledge graph · Cognitive graph · Intelligent education · Personalized learning

1 Introduction According to the NMC’s horizon 2019 report, AI is the key of higher education to revolutionize in the next four to five years [1]. As an interdisciplinary research field, with rapid development of AI, KG will support for “AI + education” in the future, and push education towards “intelligent education”. In the realization process of intelligent education, such as semantic search of knowledge, recommendation system of personalized learning, and construction of learner portrait, most of them rely on large-scale KG. From the beginning of Google’s intelligent search engine, to now, big data analytics, chatbots, personalized education and recommendation system, are all closely related to the KG. Therefore, by comprehensively analyzing the evolution of KG, it is great significance to promote innovation in the education. From the perspective of bibliometrics, this study explores the development of intelligent education by analyzing the evolution of KG.

2 Research Design 2.1 Data Collection In this study, we extracted data with “Knowledge Graph” or “Knowledge Visualization” from the Web of Science database, time range from 1996–2019. By eliminating the © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 18–24, 2020. https://doi.org/10.1007/978-3-030-63955-6_2

On the Evolution of KG Abroad and Its Application in Intelligent Education

19

irrelevant literature to KG, a total of 1210 residual literature were identified, which are used as the sample data source. 2.2 Research Methods Scientific citation data visualization analysis software (CiteSpace) and Bibliographic Items Co-occurrence Matrix Builder (BICOMB) were adopted. Through the Citation Analysis and keyword co-occurrence analysis, hot topic distribution and frontier trend of KG were discussed.

3 Development Context of KG 3.1 Chronological Analysis With the change of time, the number of literatures is an important indicator to measure the development trend of a field. In BICOMB software, “Year” is used to count the annual number of literatures, which distribution curve is significant (see Fig. 1). We can see it is to increase at a quick rate in recent years from the fitted exponential function of the curve (y = 2.043e0.1321x ) and the determination coefficient (R2 = 0.8286), the number of relating literatures.

250 200 150

y = 2.043e0.1321x R² = 0.8286

100 50 0

Fig. 1. Number of literatures on KG

The foreign research about KG began in the 1950s. In 1955, Garfield [2] pioneered the idea of using citation indexes to retrieve literatures. In 1965, Price [3] proposed the citation relationship between network and scientific literature. In 1968, J. R. Quillian proposed the semantic web. In 1977, Feigenbaum put forward the concept of knowledge engineering, and deem that knowledge engineering is the application of AI, and then ontology was introduced and became the method of representing knowledge in the real world. In 1998, Tim Berners-Lee proposed the concept of Semantic Web. On this basis,

20

Y. Zhang and B. Zhao

the specification of knowledge description on the World Wide Web was further proposed by the W3C. In 2006, Tim Berners-Lee [4] put forward the concept of linked data to generate large-scale networks based on the relation between entities. In 2012, Google proposed the “knowledge Graph”, which reflects the enhanced application of large-scale KG in intelligent search engine. The typical representatives of large-scale network knowledge, such as DBpedia [5], are built on the basis of Wikipedia structured knowledge base. With the further development of AI, it has entered the stage of cognitive intelligence, which promotes the development of construction technologies of KG.

4 Research Topic Distribution of KG 4.1 Analysis of High-Frequency Keywords The keywords of the literature are the extracted from the full text, which can objectively reflect the research hotspots in a certain field [6]. CiteSpace is used to generate the cooccurrence graphs of keywords (see Fig. 2). Co-occurrence of high-frequency keywords.

Fig. 2. Co-occurrence of high-frequency keywords

The Pathfinder Network algorithm is used to simplify the network for highlight structural features, which are mainly determined by parameter r. the relation of triangle inequality is defined as Eq. (1):  1/r  r wnk nk + 1 (1) wij ≤ k

Where wij represents the link weight between node i and node j, wnk nk+1 represents the link weight between node nk and nk+1 nodes, and r is the Minkowski distance.

On the Evolution of KG Abroad and Its Application in Intelligent Education

21

In addition to the “Knowledge Graph”, there are the top 10 high-frequency keywords: Ontology (52), Knowledge Visualization (52), Visualization (40), Semantic Web (33), Knowledge Graph Embedding (28), Knowledge Representation (21), Neural Network (16), Link Prediction (16), Dbpedia (15), Recommender System (14). As can be seen from the fluctuation of word frequency, which shows that it focuses on the key technologies of KG construction such as knowledge extraction, knowledge representation, knowledge reasoning and knowledge graph completion. 4.2 Analysis of Clustering Clustering analysis of keywords can clearly reveal the hot topics in the field, cluster was extracted from the keywords of cited literature. The extraction of cluster uses a Long-likelihood-ratio algorithm, which computed as Eq. (2):      (2) LLR = log p Cj /Vij /P Cj /Vij LLR is the log-likelihood ratio of the word Wi and Cj , the vector Vij (∝, β, γ ) is composed of the frequency (∝), concentration (β), and dispersion (γ ) of the word Wi . The vector Vij is used to determine whether Wi can be used as a feature word of category     Cj . p Cj /Vij and P C j /Vij is the density functions of the categories Cj and C j . The clustering graph of keywords as shown in Fig. 3, Modularity Q = 0.55, Mean Silhouette = 0.7309, therefore, the structure of cluster is remarkable and convincing.

Fig. 3. Keywords clustering

A total of 4 clusters were generated. They are #0 (Knowledge Visualization), #1 (Knowledge Representation), #2 (Deep Learning) and #3 (Neural Networks). knowledge visualization, knowledge representation learning, deep learning and neural network are focused on.

22

Y. Zhang and B. Zhao

4.3 On Research Hotspots of KG (1) Knowledge visualization. Knowledge visualization is a graphical method to construct and transmit complex knowledge. It uses scene visualization, relationship visualization, and process visualization to create and transfer of knowledge, in order to deepen students’ “memory” of knowledge and promote learners’ cognitive processing. The visual expression has a strong “integrity” characteristic and makes people have a “global consciousness”, which can help learners better understand and learning the knowledge. (2) Knowledge representation. Knowledge representation is the study of how to represent the knowledge of the objective world in a form that is easy for computer or machine to recognize and understand. RDF triples are mostly used to describe the relationship between entities. In recent years, with the development of deep learning, knowledge representation learning based on entities, concepts, and relationships have become mainstream [7]. Moreover, in knowledge representation, the fusions of cross-media elements and spatio-temporal dimensions are also the trend of future research. (3) Deep learning. Deep learning (DL) is a kind of machine learning based on deep neural network, which uses statistics to model specific problems in the real world and uses trained data to solve similar problems in the field. It is a semi-theory and semi-empirical model with flexible expression. DL is derived from artificial neural networks, which can be divided into convolutional neural networks (CNN) and deep belief nets (DBN). CNN is a well-known DL architecture inspired by the natural visual perception mechanism of the living creatures [8]. DBN is probabilistic generative models that are composed of multiple layers of stochastic, latent variables [9]. In addition, Google’s TensorFlow framework can used to implement open source DL systems, which can support CNN, RNN and LSTM algorithms, and also provides a tool on the web. Which can use graphics to present the real-time characteristics of the entire network. (4) Neural networks. Neural networks, often called artificial neural networks (ANN), are based on the basic principles of neural networks in biology, and are rooted in neuroscience, statistics, mathematics, and computer science. The neural network simulates the thinking of the human brain, and can have human-like understanding and judgment, which is a further extension of traditional logic calculations. As an indispensable part of machine learning, neural networks have been widely used in computer vision, decision optimization, image segmentation, cognitive science, and so on. At present, convolutional neural networks (CNN) and recurrent neural networks (RNN) are relatively mature. CNN is a neural network especially suitable for computer vision applications [10], and RNN is a neural network for processing sequence data [11]. The toolkit, such as tensorflow and sklearn has been applied neural network models to solve practical problems.

On the Evolution of KG Abroad and Its Application in Intelligent Education

23

5 Exploration of Intelligent Education 5.1 Educational KG Nowadays, the construction of educational KG is regarded as the key for intelligent education. The educational KG can be constructed to provide students with a variety of knowledge services, such as knowledge query, personalized learning path, etc. When constructing the KG about a specific domain, it is necessary to discuss with the domain experts to customize the schema of the domain KG. However, the researchers mainly focus on the coded explicit knowledge, most of which do not consider the tacit knowledge. Therefore, it is urgent to construct generic and domain KG for comprehensive construction platform. Some KG construction platforms have been constructed according to the relevant methodologies. For example, Baidu has built K-12 education KG to realize personalized learning, and providing intelligent services for students. 5.2 Cognitive Diagnosis and Cognitive Graph In terms of education and teaching, intelligent education is faced with the problem of students’ cognitive overload. As mentioned in the Cognitive Load Theory, when the cognitive load is controlled within the range of working memory, effective learning can be happened [12]. How to reduce the cognitive load and set the gradient of students’ learning are the key. However, student’s learning paths can be found and created by cognitive diagnosis and student’s cognitive load can be reduced effectively. On the other hand, according to Constructivism Learning Theory, learning is the process by which learners actively select and process external information based on their own experience [13]. In order to avoid the emergence of “Lost in learning” problem, it is urgent to build the cognitive graph based on KG through students’ cognitive diagnosis. It has “cognition” first, which can make cognitive diagnosis for students. Then “reasoning”, it is to provide a personalized learning path, which can promote students’ effective learning. 5.3 Personalized Education With development of technology such as the Internet and AI, traditional educational ideas about effective learning of students have been inherited and surpassed. David Pawl Ausubel believes that students’ learning should be as meaningful as possible if it is valuable [14]. Therefore, the development of personalized education should also aim at intentional, active, real, constructive and cooperative learning. It is urgent to create a fair, open and personalized education environment by the deep combination of AI and education, which was proposed at the international conference on AI & education in May 2019. In particular, with the coming of 5G era, unstructured data transmission can be achieved through the 5G’s wireless communication technology, which provides a more three-dimensional digital environment for intelligent education [15]. In the future, with the progress of KG and 5G, intelligent education will become the main way of online learning.

24

Y. Zhang and B. Zhao

Acknowledgments. The research is supported by a National Nature Science Fund Project (No. 61967015), and Undergraduate Education and Teaching Reform Project in Colleges and Universities of Yunnan (JG2018060).

References 1. Alexander, B., Ashford-Rowe, K., et al.: EDUCAUSE Horizon Report: 2019 Higher Education Edition. Louisville, Co.: EDUCAUSE. Accessed 11 Jan 2020 2. Garfield, E.: Citation indexes for science: a new dimension in documentation through association of ideas. Science 122(3159), 108–111 (1955) 3. De Solla Price, D.J.: Networks of scientific papers. Science 149(3683), 510–515 (1965) 4. Berners-Lee T.: Linked data-designissues. http://www.w3.org/DesignIssues/LinkedData. html. Accessed 12 Jan 2020 5. Auer, S., Bizer, C., Kobilarov, G., Lehmann, J., Cyganiak, R., Ives, Z.: DBpedia: a nucleus for a web of open data. In: Aberer, K., et al. (eds.) ASWC/ISWC - 2007. LNCS, vol. 4825, pp. 722–735. Springer, Heidelberg (2007). https://doi.org/10.1007/978-3-540-76298-0_52 6. Wang, X.M.: Research on the teaching quality of colleges and universities in China: track, hotspots and future trend – a Citespace visualization based on fourteen core journals of higher education. Educ. Mon. 1, 91–103 (2018) 7. Liu, Z.Y., Sun, M.S., Lin, Y.K., et al.: Knowledge representation learning: a review. J. Comput. Res. Develop. 53(2), 247–261 (2016) 8. Gu, J.X., Wang, Z., et al.: Recent advances in convolutional neural networks. Pattern Recogn. 77, 354–377 (2018) 9. Hinton, G.: E.: Deep belief networks. Scholarpedia 4(6), 5947 (2009) 10. Zeiler, Matthew D., Fergus, Rob: Visualizing and understanding convolutional networks. In: Fleet, David, Pajdla, Tomas, Schiele, Bernt, Tuytelaars, Tinne (eds.) ECCV 2014. LNCS, vol. 8689, pp. 818–833. Springer, Cham (2014). https://doi.org/10.1007/978-3-319-10590-1_53 11. Graves, A., Mohamed, A.R., Hinton, G.: Speech recognition with deep recurrent neural networks. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, pp. 6645–6649 (2013) 12. Paas, F., Renkl, A., Sweller, J.: Cognitive load theory: instructional implications of the interaction between information structures and cognitive architecture. Learn. Sci. 32(1/2), 1–8 (2004) 13. Ültanir, E.: An epistemological glance at the constructivist approach: constructivist learning in Dewey, Piaget, and Montessori. Int. J. Instr. 5(2), 195–212 (2012) 14. Ausubel, D.P., Cook, H.: Educational psychology: a cognitive view. Am. J. Psychol. 83(2), 303 (1970) 15. Lewis, J.A.: How 5G will Shape Innovation and Security: A Primer. CSIS, Washington (2018)

Effectiveness Evaluation of Network Security Knowledge Training Based on Machine Learning Quan-wei Sheng(B) Changsha Medical University, Changsha 410219, China [email protected]

Abstract. Due to the problem of inaccurate evaluation results of traditional methods for evaluating the effectiveness of network security knowledge training, a method for evaluating the effectiveness of network security knowledge training based on machine learning is designed. First, establish an evaluation index for the effectiveness of cybersecurity knowledge training, and then develop an evaluation standard for the effectiveness of cybersecurity knowledge training. The standard performance status is the teaching effect that employees should achieve after training. Finally, calculate the weight of each indicator to complete the evaluation of the effectiveness of cybersecurity knowledge training. The experimental comparison results show that the effectiveness evaluation method of network security knowledge training based on machine learning designed this time is more accurate than traditional methods, and has practical significance. Keywords: Machine learning · Network security knowledge · Training effectiveness · Indicators · Weights

1 Introduction In recent years, our country pay more and more attention to in network security knowledge training work, and the original network and the state economic and trade commission and other departments have issued a series of laws, regulations and safety production training, carry out a series of safety training is given priority to with safety publicity and education activities, to establish the occupational safety and health training center and network security knowledge training center, carried out various areas, various industries and at all levels, all kinds of safety production training, especially the business operators and managers qualification training, safety supervision inspectors work safety qualification training, new workers safety education and special operations personnel safety operation qualification training, Initially formed a certain scale of safety production training network.

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 25–37, 2020. https://doi.org/10.1007/978-3-030-63955-6_3

26

Q. Sheng

Cyber security knowledge training is an important way to increase trainees’ security knowledge, improve security skills, and improve security concepts, thereby improving their security qualities, in order to prevent accidents and increase the level of intrinsic safety of human resources in enterprises. The evaluation of training effectiveness directly affects the realization of safety goals. Under the influence of the above factors, the evaluation of the effectiveness of safety training in China has been in an inefficient stage for a long time, and the employees’ unsafe behaviors are still not tangible. Although the current state has increased investment in human and material resources for the evaluation of safety training work, which has made certain achievements in China’s safety training work, because the evaluation of the effectiveness of safety training in our country is still in its infancy, it needs continuous supplement And perfect, a method for evaluating the effectiveness of network security knowledge training based on machine learning is designed. Machine learning is a multi-disciplinary and interdisciplinary discipline that involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. Specialize in how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance.

2 Establishment of Evaluation Index of Network Security Knowledge Training Effect Traditional network security knowledge training effect evaluation path is mainly divided into two categories, one is the internal training effect evaluation [1], 2 it is after the internal training will prepare the good evaluation questionnaires distributed to the trainees, let trainees to training instructors, training course content, and to evaluate training organization support, and so on and so forth, and according to the answer of the trainees, formulate improvement measures. However, the evaluation index of this method is relatively single. Aiming at the existing problem of the traditional method, the safety of the staff training effect evaluation system mainly involves three kinds of object, namely the security training management personnel, mainly be responsible for the training plan formulation and implementation of professionals, practitioners is direct embodiment of training do you have any good results, improve the work ability of employees is the ultimate goal of network security training, trainer is imparting knowledge and skills, the trainer’s personal ability and behavior directly influence on the capability and the extent of the trainees to accept. According to the different characteristics, requirements and processes of these three kinds of objects, different methods are adopted for analysis. Since the development of training plan involves many evaluation indicators, a wide range, a long periodicity of the process, and the role and importance of each step [2] are not consistent, the machine learning evaluation method is determined for evaluation as shown in the following figure (Fig. 1):

Effectiveness Evaluation of Network Security Knowledge Training

27

internal training

internal training

Questionnaire

internal training

internal training

Trained employees

internal training

Training file management

Fig. 1. Network security knowledge effect evaluation subject

Summarize the training work of the company as required, and write the evaluation report. So as to grasp the training situation and make corresponding adjustment measures according to the evaluation report. Based on the above analysis, the evaluation index [3] is established as follows (Table 1):

28

Q. Sheng Table 1. Evaluation index of network security knowledge effect

First-level indicators

Secondary indicators

Tertiary indicators

Evaluation of training organizations

Organization support

Attention of training opinions Notification in a timely manner Logistics assistance, complete equipment

Overall evaluation

Expectations before training Post-training evaluation Willing to participate again

Evaluation of training courses

Course quality

Theoretical level content Practice level content Arrangement and logical levels of course content

Trainer Evaluation

Curriculum, job relevance

Training courses for work

Personal image

Manner and behavior Temperament Personal affinity

Sense of responsibility

Accurate and reasonable teaching logic arrangement Degree of positive response from trainees Able to arrange course content accurately and reasonably Preparation of teaching materials Can inspire trainees

Trainee evaluation

Training experience summary

3 Effective Evaluation Criteria for Network Security Knowledge Training Standard performance status is the teaching effect that employees should achieve after training, that is, the knowledge, skills, and attitude that employees should have after training [4], which is mainly based on employee competence. It should be noted that the competency index is not the same as the training effect evaluation index. The competency index mainly refers to the performance status of the employee’s work, and the training effect evaluation index refers to the training project itself. In some aspects, there is a certain relationship between competency indicators and training effectiveness evaluation indicators. For example, training for a specific training project may be to improve a certain skill or ability of employees, but not all competencies. Evaluation needs to include both a certain skill part and other parts, such as learning situation and organizational

Effectiveness Evaluation of Network Security Knowledge Training

29

improvement situation. Since there are many content evaluations of the effectiveness of cyber security knowledge training, it is understood that the performance of the excellent charging staff in the respondents’ minds on the three dimensions of knowledge, skills, and attitudes [5] has been transformed into measurable standards. After in-depth interviews, combined with the job description, the following knowledge, skills, and attitude requirements of the fee-charging staff are summarized, as shown in the following table (Table 2): Table 2. Network security knowledge training effect evaluation criteria Serial number First dimension Secondary dimension 1

Know how

General business knowledge Industry knowledge Laws and regulations Civilized service requirements

2

Skill

Communication skills Learning ability Special event handling capabilities Emotional control Computer Skills Equipment operation capability

3

Attitude

Honesty Service awareness Safety consciousness Professionalism

As a tool for performance improvement, training is often a short-term behavior. In other words, the effect of training on employees mainly focuses on the improvement of knowledge, skills and part of the attitude, while certain personality qualities cannot be acquired through short-term training [6]. Based on this, “resilience” was excluded from the measurement of actual performance of employees in this study. In the decision-making of qualitative problems, AHP is an excellent method, which is based on the comparison of two evaluation objects and the construction of judgment matrix with the comparison results: ⎡

⎤ 1 1/2 1/4 1/4 ⎢ 2 1 1/2 1/3 ⎥ ⎥ A−B=⎢ ⎣ 4 2 1 1/2 ⎦ 4 3 2 1

(1)

30

Q. Sheng

In the common formula (1), A and B respectively represent matrix elements [7]. Since the judgment matrix is the basis for determining the weights, the matrix needs to meet the requirements of consistency. Therefore, the errors and compatibility of the judgment matrix must be analyzed, Which is calculated as: C.R =

yo Uf

(2)

In formula (2), C.R represents the consistency test result, Uf represents the weight vector of the matrix, and yo represents the index system containing the weight. Then, the judgment matrix is normalized, and the calculation formula is: d= 

r t∗m

(3)

i=1

In formula (3), d represents the normalized processing parameter of the index,



t

i=1

represents the level of the evaluation index, m represents the whitening weight function of the evaluation gray class, and r represents the gray evaluation coefficient. Finally, for the consistency check of the evaluation matrix, its calculation formula is: CU =

ηc − nm N

(4)

In formula (4), CU represents the matrix consistency index, N represents the maximum characteristic root of the matrix, ηc represents the matrix element, and nm is the weight coefficient of each layer of the index. According to the above process, the calculation of the index weight [8] is completed, and the evaluation criteria for the effectiveness of network security knowledge training are determined according to the determined weight index.

4 Realization of Effectiveness Evaluation of Cyber Security Knowledge Training Based on the establishment of the above-mentioned network security knowledge training effectiveness evaluation indicators and the establishment of the network security knowledge training effectiveness evaluation standards, the network security knowledge training effectiveness evaluation [9], the overall evaluation structure is shown below (Fig. 2):

Effectiveness Evaluation of Network Security Knowledge Training

31

Evaluation Keep the situation

Implementation status

aims

Expectation

result

Original condition

Fig. 2. Evaluation structure of network security knowledge training

In the scientific research of various fields, it is often necessary to observe a large number of variables reflecting things and collect a large number of data in order to analyze and find rules. Multi variable [9] large samples will undoubtedly provide rich information for scientific research, but to a certain extent, it also increases the workload of data collection, more importantly, it increases the complexity of problem analysis. Because there is a certain correlation between the variables, it is possible to use less comprehensive indicators to analyze all kinds of information in each variable, and the comprehensive indicators are not related to each other, that is, the information represented by each indicator does not overlap the comprehensive indicators, which not only retains the main information of the original variables, but also has some superior properties than the original variables, which makes it possible to study complex problems It is easier to grasp the main contradiction. In this way, the comprehensive indicators can be named according to professional knowledge and the unique meaning reflected by the indicators. This method is called factor analysis, and the comprehensive index representing all kinds of information is called factor or principal component. According to the purpose of factor analysis, we know that the comprehensive indicators should be less than the original variables, but the information contained should be relatively less loss, so the

32

Q. Sheng

relationship between each factor and the original variables can be expressed as: e x = fg +

d

(5)

i=1

In formula (5), fg represents the original variable vector,

e

d represents the common

i=1

factor load coefficient and residual, and x represents the principal component vector. Based on this, the comprehensive evaluation value is calculated, and the control of the safety training process is only good. Generally, more efforts need to be made at all levels to further improve the control level of each process. When there are several implementation schemes participating in the evaluation jointly, the company also They can be ranked and selected according to their comprehensive evaluation value, and its calculation formula is: o (6) F = ∗ jk D In formula (6), F represents the weight vector of each criterion layer, D represents the comprehensive evaluation value, o represents the influence parameter, and jk represents the degree of correlation between the indicators. According to the above process, combined with the above indicator system to establish an evaluation hierarchy, as shown below (Fig. 3): Based on the above, a comprehensive evaluation is carried out to achieve scientific decision-making, continuously improve the business level of managers in safety training, and then improve the training effect of enterprises, reduce the occurrence of accidents, and enhance the overall safety level of the enterprise.

5 Experimental Comparison In order to verify the effectiveness of the evaluation method of network security knowledge training based on machine learning designed this time, an experiment was conducted, and the traditional method was compared with the design method to compare the accuracy of the two methods. 5.1 Experimental Scheme This project training is for 5 website personnel of a company. A total of 27 applicants applied for project training, of which 24 are eligible for registration. Therefore, this study evaluated 24 trainees. Specifically, one day after training for each training course, the course is assessed by taking a test of the course, the results are recorded in the registration form, and the training center evaluates the enthusiasm of the trainees. The center distributed behavior-level questionnaires to 24 trainees, and recovered 24 valid questionnaires, with a questionnaire recovery rate of 100%, and the supervisors of the operation department evaluated the behaviors of trainees in each department after training. Among the valid samples, 11 were male (46%), 13 were female (54%), the average

Effectiveness Evaluation of Network Security Knowledge Training

33

Personal image

Lecturer situation

Sense of responsibility Teaching quality

Course content Subject fit Reaction layer Organization support

Content availability

Content depth

Initiative Satisfaction Learning layer knowledge level

Arrange satisfaction Environmental satisfaction

Skill levels

Behavioral layer

Professionalism

Work efficiency

teamwork Result layer Turnover rate

Fig. 3. Hierarchical structure of comprehensive indicators

age was 29 years, the average working age was 6 years, 63.3% had a college degree or below, and undergraduates and above accounted for 26.7%. Sort out data from 24 valid questionnaires. After the training, the trainees were asked to solve the problem of network attack. In the experiment, each port of the target server was requested to connect in turn, and then some open ports were found. At this time, cyclic neural network is used to learn the process. After learning, the neural network will be able to predict whether the next operation is normal access or network attack according to the characteristics of the current data. The scanning target is as follows (Fig. 4):

34

Q. Sheng

Fig. 4. Content of network security training

After the training of the experimental personnel, the above network security events are solved, and the data captured by the server is as follows (Fig. 5):

Fig. 5. Server capture data

Then evaluate the effectiveness of network security knowledge training. There are many data contents involved in the evaluation process. An experimental platform is established for this purpose, as shown in the following figure (Fig. 6): The experiment was carried out with the above platform, and the specific experimental results are shown as follows.

Effectiveness Evaluation of Network Security Knowledge Training

Data register

35

Client

Address register

the Internet

Memory RAM

application

Server control unit

Fig. 6. Experimental platform setup

5.2 Analysis of Experimental Results The comparison between the traditional evaluation method and the design evaluation method is shown in the following figure (Fig. 7): Analysis of the above comparison results shows that the accuracy of the design method in the training effect evaluation is high, up to 98%, while the accuracy of the traditional method in the training effect is poor, up to 50%. In the five experiments, the accuracy of the traditional method in the training effect is lower than the design method, because the design method can fully analyze the evaluation of network security knowledge Standard, as well as the corresponding index weight, thus ensuring a higher evaluation accuracy. Therefore, through the above experiments, it can be proved that the design method has practical application significance and higher evaluation accuracy than the traditional method.

36

Q. Sheng

Evaluation accuracy /%

100 80 60 40 20 0 1

2

3

4

5

Experiments / times Design method this time Traditional method

Fig. 7. Comparison of training effect evaluation

6 Concluding Remarks How to objectively and effectively establish the training effect evaluation index system to evaluate the enterprise training effect, so as to diagnose and modify the enterprise training program, and then improve the effectiveness of the training program, and promote the healthy development of the enterprise, is an urgent problem to be solved by the enterprise, is also worth the attention of the academic community. The traditional evaluation method is of low accuracy, so a new evaluation method of network security knowledge training based on machine learning is designed, and the effectiveness of this method is proved. To sum up, the method of this study has the following characteristics: applying performance technology to the evaluation system of enterprise training effect. Previous studies on the evaluation system of enterprise training effect by scholars mostly focus on simple problem analysis or overview, but lack systematic and targeted analysis. This article applies machine learning to the diagnosis of training effect evaluation problems, making the problems more obvious and the causes more targeted. The method of fusion of quantitative and qualitative analysis is used for research. Quantitative analysis and qualitative analysis are softened together, making the index construction process more standardized. With a certain degree of reliability. Even so, this study still has certain limitations: in the selection of evaluation indicators, this study uses as much independence, efficiency and operability as possible; for result-level indicators, vehicle toll revenue and employee turnover rate They will be affected by many factors, and the changes may not only be caused by training, but also by other factors. However, although the changes in this research indicator take this into account, in the evaluation of the examples, the influence of other factors cannot be controlled.

Effectiveness Evaluation of Network Security Knowledge Training

37

References 1. Cheng, Z., Chao, K.: Research on virtual and real combination system for training in mechanical safety protection. Chin. Saf. Sci. J. 28(5), 166–171 (2018) 2. Bingfeng, H., Qianli, Y., Xizhong, Y., et al.: Evaluation of training course of hepatic disease in grass roots clinical hepatic physicians in some poverty-stricken counties in Shanxi and Shaanxi provinces. Chin. J. Epidemiol. 39(5), 636–639 (2018) 3. Huan, Z., Xiaotian, C., Xuejun, Z., et al.: An analysis of residents’ self-evaluation and facultyevaluation in internal medicine standardized residency training program using Milestones evaluation system. Chin. J. Intern. Med. 57(6), 440–445 (2018) 4. Xincheng, W., Jide, S., Zhaopu, Z., et al.: A model for game between safety training provision of firm and worker’s attendance to it in construction industry. Chin. Saf. Sci. J. 28(7), 159–164 (2018) 5. Yi, W., Zhenqun, L., Sicheng, X.: Causative analysis on factors affecting expansion effect of mechanical ventilation technology in Xinjiang primary hospital. Chin. J. Integr. Tradit. West. Med. Intensive Crit. Care 25(5), 546–548 (2018) 6. Jing, W., Xiaoqing, J., Ruining, L., et al.: A study on the effect of CPR training with intensive training and real-time feedback device on teaching effect. Chin. J. Emerg. Med. 28(2), 199–202 (2019) 7. Huiliang, W., Kaifa, C., Yunfei, L., et al.: Water regulation measures and effect evaluation of central Henan urban agglomeration based on artificial neural network. Yellow River 41(6), 58–61 (2019) 8. Yanye, D., Yingying, J., Jianyu, W., et al.: Assessment of effect of situated learning on developing crisis intervention skills. J. Shanghai Jiaotong Univ. (Med. Sci.) 39(5), 539–543 (2019) 9. Yiwei, F., Shengran, W., Yuyu, S., et al.: Effectiveness evaluation of WeChat combined with traditional missionary methods in college AIDS health education. Chin. J. Sch. Health 40(4), 598–601 (2019)

Based on the Big Data Program Language Learning Website Generation System Quanwei Sheng(B) Changsha Medical University, Changsha 410219, China [email protected]

Abstract. In order to improve the learning efficiency of students, a website generation system based on big data is proposed. Under the background of big data, the hardware design of the system is completed by using the overall architecture design of the learning website generation system and the design of the learning website generation server. Through the collection of program language learning website data, the design of program language learning website generation database and the design of program language learning website generation process, the software design of the system is completed and the generation of program language learning website is realized. The test results show that compared with the website generation system based on cloud computing, the students’ learning efficiency is higher. Keywords: Big data · Programming language · Learning website · Generation system

1 Preface With the rapid development of electronic technology and network technology, today’s society has become an information society. The new talents required by the information society must have the ability to acquire, analyze and process information, as well as practical ability and innovative spirit. Traditional programming language teaching materials and classroom teaching have limited the improvement and comprehensive development of students’ information literacy to a certain extent [1]. Because of this, the Ministry of Education stated in the “Notice on Popularizing Information Technology Education” that from 2001, it took five to 10 years for China to popularize information technology education in primary and secondary schools, use information technology to drive education modernization, and work hard to achieve China. Leaping development of basic education [2]. At present, information technology courses in middle schools have become subjects for senior high school examinations. It can be seen that information technology courses have become more and more important in subject teaching, so higher requirements have been placed on the teaching of procedural language learning courses. Many schools have begun to pay attention to the teaching of information technology courses, but the author found in teaching practice that because the basic knowledge of © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 38–53, 2020. https://doi.org/10.1007/978-3-030-63955-6_4

Based on the Big Data Program Language Learning Website

39

programming language and information technology provided in textbooks is relatively boring, operation exercises are not attractive to students, and it is difficult to cause this problem. The learning interest of students at the age stage results in low learning efficiency. At the same time, due to the rapid development of information technology, there is a lot of knowledge update in the textbooks that cannot keep up with the times and is outdated and cannot meet the needs of students. Therefore, the teaching methods and learning methods of traditional procedural language learning websites can no longer meet the needs of students’ comprehensive development, and new teaching models are urgently needed to promote middle school information technology courses to meet the requirements of the times [3]. The traditional meaning of the program language learning website generation system, is each program language independent design, separate coding, development workload, repeated coding more, with the increase in the demand of the program language learning website, professional developers are difficult to deal with the development of the website. In order to benefit more teachers and students, and considering the structural similarities of program language learning websites, it is imperative to develop high-quality program language learning websites quickly and efficiently. How in a short period of time to develop high quality programming language learning website, just rely on the development of professional and technical personnel can’t be meet a lot of demand, which makes programming language learning website generation system become the research topic, how to design a kind of easy to use this website generation system, to help don’t know the web site development and design technology of teachers can independently easily, quickly to learn architecture framework, design programming language learning website, with independent personality has important practical significance. The research of this topic is based on solving the practical problems in education and teaching from the technical level, aiming to play the purpose of technology serving for teaching. In terms of research, the combination of investigation and analysis, theory and practice, and demand and application will help enrich teachers’ teaching methods and achieve good learning results. Design a big data-based program language learning website generation system, which not only saves the work of web page design and production that needs the participation of technical personnel, but also satisfies the designer’s demand for page personalization through a variety of templates. The research results will be helpful to promote the development of distance education and realize the goal of life-long education, which has considerable promotion value.

2 Programming Language Learning Website Generation System Hardware Design 2.1 Overall System Architecture Design Establish a convenient and easy-to-use big data-based programming language learning website generation system, so that teachers can quickly build the framework page of the learning website without the help of technical staff, thereby focusing on designing teaching content pages. At any time, teachers can change the layout of the website by changing the website template or file [4]. As technicians in the educational technology

40

Q. Sheng

department, we will make full use of the web design experience accumulated in our work, and constantly design and add learning website templates for the system, making the system increasingly practical. According to the goal of the system, this paper analyzes the functional requirements of the website generation system of program language learning, which is summarized as follows: (1) Requirements for information collection Provide information collection interface to fill or modify the program language learning website information: Information collection is carried out in a wizard-style filling form, which is simple and easy to use; the information collection page is also an information display page. Limit the information submitted to ensure the validity of the data and provide secure data storage. The collected information includes configuration information of program language learning website, directory information of three levels, template selection information, etc. (2) the framework generates requirements This part is the most important functional requirement of this system. According to the collected information, each frame file of the learning website is generated on the server side, and a folder for placing the content page file is generated. The directory information should be presented in a tree structure in the frame page. (3) Template requirements The system provides template management functions, including importing templates, modifying templates, deleting templates, etc. Develop flexible template design rules to enable template designers to design template pages with diverse structures. (4) Interaction requirements In the first - level directory to provide optional learning site commonly used interactive modules, including answering module, discussion module, homework module. (5) User management requirements Perform user authentication. For system security, this system does not allow two people to log in with one account at the same time. The password is irreversibly encrypted to effectively ensure the security of the system and user data. According to the analysis of system requirements, the website generation system of program language learning is divided into presentation layer, business logic layer and database layer [5]. The presentation layer is responsible for providing a complete and unified interactive interface for users, the business logic layer is responsible for the processing of core functions, and the database layer is responsible for the access to database data. The overall architecture of the system is shown in Fig. 1.

Based on the Big Data Program Language Learning Website Presentation layer

Business logic layer

Teacher

Database layer

System information base A tool for building program language learning websites

Student

41

user interfac e

User information base Program language library

Template library system management tool database

Administrators

Client machine

Web server

database server

Fig. 1. Overall architecture of the system

B/S structure is adopted in the whole website generation system of programming language learning Browser server architecture, system users (teachers, students, system administrators) log in to the system through the browser, put forward the business request of system management and program language learning website construction to the system through the user interface of the presentation layer, the system business logic layer processes the business, and operates the database when necessary. Under the B/S structure, the management, resource allocation, database operation, business logic management and dynamic loading of the whole system are centralized on the server, which is easy to deploy and manage. As can be seen from the overall system architecture diagram, the design of the system ultimately rests on the user interface, system functions, and database. 2.2 Programming Language Learning Website Generation Server Design The programming language learning website generation system is based on the B/S architecture and is divided into two parts, client and server. The server uses the RT5350 chip equipped with the Linux operating system as the development platform [6]. It mainly completes the functions of data acquisition, compression, and network transmission of the programming language learning website. The overall design structure of the server is shown in Fig. 2:

42

Q. Sheng USB camera

data acquisition

data compression

network transmission

user

Fig. 2. Overall design structure of server side

From the above figure, it can be seen that the server side of the website of program language learning USES the camera of USB interface as the data acquisition device of the website of program language learning, and completes the data collection function of the website of program language learning by calling the V4L2 function interface. Program language learning website data compression using the JPEG compression algorithm introduced above, through the transplantation of Libjpeg library to achieve. In the network transmission part, mjpg-streamer is transplanted to the streaming media server, and Socket communication is used to complete the wireless network transmission of the website data of program language learning. Through the above operations, we have compressed the collected program language learning website data, and the server listens to the client’s access request after the collection. If an access request is received from the client, the server returns the program language learning website data to the client. This system is based on the TCP transmission protocol, and uses Socket communication to realize the program language learning website data transmission. The working structure of the programming language learning website generation server is shown in Fig. 3.

Based on the Big Data Program Language Learning Website

43

Fig. 3. Working structure of the server for generating program language learning website

As can be seen from the above figure, the programming language learning website generation server first creates a Socket socket through the Socket () method and returns a descriptor of the Socket. The Bind () method is then used to bind the IP address and port number to the programming language learning website to generate a server-side binding. Call the Listen () function to listen to whether the client sends a request to the server. At this point, the client creates a Socket object and returns a descriptor of the Socket [7]. And call the Connect () function to send a connection request to the server through the IP address and port number, and the programming language learning website generates a server to receive the connection request from the client through the Accept () function and establish a connection. At this point, the Socket connection has been established, and the two communication parties can send data to each other. After communication is complete, close the Socket connection.

3 Program Language Learning Website Generation System Software Design 3.1 Collecting Program Language Learning Website Data In order to improve the learning efficiency of students, it is necessary to compress the data before collecting the website data. Due to the large amount of original data

44

Q. Sheng

collected by the generation server of the program language learning website, it can not be smoothly transmitted in most networks. Therefore, we need to compress the collected website data [8]. Using JPEG as the data encoding method of the system, the advantages of JPEG encoding are its high compression ratio, simple algorithm, low requirements for hardware, which is very suitable for real-time transmission of program language through the network to learn website data. The compression structure of program language learning website data is shown in Fig. 4.

Fig. 4. Data compression structure diagram of programming language learning website

The compression steps of the program language learning website data are as follows: Step 1: Discrete cosine transform Discrete cosine transform is often used in signal processing and data processing for lossy data compression of signals and data. After the big data conversion, each data will generate a DC coefficient and 63 AC coefficients. The main purpose of using DCT to convert spatial domain data to frequency domain data is to focus the largest and most important information in a programming language learning website on a small amount of data. The formula for the DCT transformation is as follows: F(u, v) =

1 (2i + 1)uπ (2j + 1)vπ C(u)C(v) cos cos 4 16 16

(1)

Based on the Big Data Program Language Learning Website

45

Step 2: Quantitative After DCT conversion, the data memory block is processed to make each coefficient value smaller, so as to obtain better compression effect in subsequent coding. For low frequency programming language learning website data quality most value and quantitative method, DC coefficients using the most exquisite processing after the reduction of almost no distortion, numerical data quality to affect language learning program site high frequency coefficient, the least important of JPEG is used the most rough quantitative way to quantify, so can get better compression ratio. Step 3: Encoding After quantization, DC code words and AC run code words are obtained. The DC codeword is coded using a differential coding (Differential Pulse Code Modulation: DPCM) method, while the AC codeword code is sequentially coded using a run-length coding method. In order to further improve the compression ratio, Huffman coding is also required for encoding. The first thing the system needs to realize is the collection of data from the program language learning website, and the data collection interface is also the data display interface of the original system [9]. The system USES multiple groups of one and two dimensional arrays to temporarily store the data submitted by the user. The array name is saved in the form value, and the form name corresponds to the array name to realize the data transfer and data storage between pages. In a wizard-style way, each page provides “previous step” and “next step” buttons to guide users to enter information. Its functional flow is shown in Fig. 5 below.

46

Q. Sheng start Defining variables and arrays Read basic information table and primary directory table information Display and enter basic website information Save site data Display and input level 1 catalog information

Next step

N

Y Save site data Display and enter secondary catalog information N

Next step Y Save site data

Display and enter three-level catalog information

Next step

N

Y Select template

N

Next step Y Save site data End

Fig. 5. Data collection flow chart of program language learning website

In the programming language learning website generation system, you need to complete the collection of programming language learning website data through the V4L2 interface. V4L2 (VideoFor Linux 2) is a new framework for developing acquisition device drivers under Linux. It can provide a standard API interface for application development. During the development process, developers can directly use the function interface provided by V4L2 to complete the development of the driver, which greatly reduces the workload of driver development. 3.2 Design of Database Generated for Website of Program Language Learning To build a good data organization structure and database, so that the whole system can quickly, conveniently and accurately call and manage the required data, is an important indicator to measure the quality of system development.

Based on the Big Data Program Language Learning Website

47

In the design of database system, the following factors should be considered: The database must be clearly structured and reasonably arranged. The data structure of the database must follow the principle of normalization and standardization to ensure the data exchange of the previous program and the normal operation of the system. When designing data tables, on the one hand, data redundancy should be minimized, storage space should be fully utilized, and the possibility of data consistency problems should be reduced. On the other hand, appropriate redundancy should be considered to improve operation speed and reduce development difficulty. Must ensure the correctness and consistency of data during the exchange process. Pay attention to database security. According to the analysis of the data relationship, the data tables mainly designed and used by this system include the basic information table, catalog information table, template information table, and user information table of the programming language learning website [10]. The design of the data table is shown in Tables 1, 2, 3 and 4. Table 1. Basic information table of programming language learning website Name

Data type

Size

Is it empty?

SiteNameC

Text

50

Yes

SiteNameE

Text

50

Yes

SiteNameF

Text

50

Yes

SiteTitle

Text

50

Yes

InstallDir

Text

30

Yes

WebmasterEmail

Text

100

Yes

Copyright

Remarks



No

GuestBook

Yes/no



No

Bbs

Yes/no



No

Exercise

Yes/no



No

The database design of the programming language learning website can provide query and storage processing of website data for the programming language learning website.

48

Q. Sheng Table 2. Catalogue information table of programming language learning websites Name

Data type

Size Is it empty?

Id

Auto number





MenuName Text

50

No

Href

Text

50

Yes

Down

Yes/no



No

Iframe

Text

50

Yes

Directry

Text

50

Yes

sequence

Double precision type –

No

Table 3. Programming language learning website template information table Name

Data type

Size

Is it empty?

templateID

Auto number





Templatename

Text

50

No

Showfilename

Text

50

Yes

Css

Remarks



Yes

mainHtml

Remarks



Yes

Indehtml

Remarks



Yes

Menu2

Remarks



Yes

selected

Yes/no



Yes

Table 4. User information table of programming language learning website Name

Data type Size

Is it empty?

ID

Auto number





AdminName

Text

50

Yes

Password

Text

50

No

Purview

Long integer



No

LastLoginTime

Yes/no



Yes

LastLoginTime

Yes/no



Yes

LoginTimes

Long integer



Yes

Based on the Big Data Program Language Learning Website

49

3.3 Programming Language Learning Website Generation Process Design After the user enters the system, the system first provides a number of information entry windows for the user. The first is a basic information entry window for entering basic information such as the Chinese name, English name, title, copyright, and administrator mailbox of the website. The user clicks “Next”, the system will enter the second entry window for entering the first-level directory information, followed by the second-level directory information entry window and the third-level directory information entry window. The entered directory information includes the directory name, whether there is Sub-level, link file name, folder name where the content page is stored, etc. The last entry window is the template selection window. The name, introduction, and effect map of each template are placed in the window for the user to make a choice. The program language learning site generation process is shown in Fig. 6. Start

Choose a programming language learning website

instructional design

Prepare program language resources

Building the knowledge structure of program language

Carry out a variety of learning evaluation activities

End

Fig. 6. Flow chart of programming language learning website generation

Choose a programming language learning website As mentioned earlier, since the information technology discipline is different from other disciplines, the information technology knowledge is constantly developing. Therefore, to use the programming language learning website to carry out teaching activities, when making a programming language learning website, you must first select topics and choose the questions should be derived from the needs of education and teaching, and at the same time they must be expandable and practical.

50

Q. Sheng

Teaching design When choosing a programming language, teaching design, be sure about the basic level of operating environment, through the programming language study and research to complete the teaching goal, the knowledge content to a framework is presented to students, how to design what learning activities, use what kind of teaching strategy, need what kind of teaching resources, can be collected through what way, use what way to test the students can understand and grasp how much knowledge through learning. “Application” programming language teaching goal is to make students through the program language to understand and master the basic operation of office software, programming language knowledge is divided into introduction, skills, art appreciation and so on several parts to present knowledge in the program to students in the process of language learning website generated by computer operations training their own ability, to master the knowledge of programming languages. The resources used can be directly used in the library of programming language materials, can also be collected through the Internet, and the relevant part of the textbook can be directly used in the content of the textbook. Preparation of program language resources The preparation of programming language resources can be synchronized with the teaching design. After the programming language knowledge structure is basically determined, the corresponding materials and materials can be prepared. The materials in the resource library can be used directly, or the Internet, CD-ROM, newspaper digest or textbooks can be collected, and you can make your own materials if you can. In the preparation of programming language resources, attention should be paid to the normative nature of media formats and their relevance to the platform. When adding resources to the resource library, you should classify them according to the knowledge structure, in order to manage the resources and use them in the process of making a programming language learning website. Building the knowledge structure of program language Click “program language knowledge” to directly enter the program language knowledge module. After using “modify column” to edit the basic part of the program language learning website, the whole website generation process is what you see is what you get. After building the knowledge structure of programming language, you can start to edit a single page. You can edit the format of text paragraphs and insert multimedia materials into the page directly through the online editor of the system. Carry out a variety of learning evaluation activities The online test can be realized through the program language test module. The test paper can be generated from the topic selection of the test bank or uploaded by the students themselves. The students can achieve self-evaluation after the test. Through the work submission module and the forum can realize the teacher to the student’s evaluation and the student’s mutual evaluation. Through this program language, we can see that using the generating system can greatly simplify the work of teachers, to avoid the network technology in the process of network teaching to carry out the obstacles, can effectively improve the efficiency of programming language learning website development, subject teachers can be free, so that they will focus on the organization of the teaching design and teaching content.

Based on the Big Data Program Language Learning Website

51

In conclusion, in the context of big data, the hardware design of the system is completed by using the overall architecture design of the system and the design of the learning website generation server; the software design of the system is completed by collecting the data of the learning website, designing the data base of the learning website and designing the generation process of the learning website, The program language learning website is generated.

4 System Test 4.1 Test Environment Considering the need to test the compatibility of the system, the environment used in the test process is as follows: Processor: Intel core 1.83G dual-core Memory: 1 GB Operating system: Windows XP HOME Browser: IE7.0 Because the system is a platform for learners to study, discuss and test online, learners need to use a browser to use this platform. They must use the current mainstream browsers on the market to be able to browse to the page of the website and implement related Features. 4.2 Test Methods and Procedures In order to verify the practicability of the application language learning website generation system, the application language learning website generation system based on cloud computing is adopted as the experimental comparison object. The specific operation steps of the experiment are as follows: Step 1: run the install.asp file of the system on Internet explorer. Step 2: after entering the following basic information of the website in the pop-up window, click “next”. Chinese name: computer network English name: computer network Website subtitle: project 151 Website title: computer network learning website Email address of administrator: [email protected] Step 3: in addition to the two interactive modules of “message” and “forum”, there are three first-level directories on the website. Therefore, after selecting “message” and “forum”, enter the number “3” and click “next”. Step 4: after entering the level 1 directory information in Table 5, click next. Step 5: after entering the information of the following secondary directory, click “next”. Step 6: after entering the following three-level directory information, click “next”. Step 7: select a template and click next. Step 8: click “ok” and the system will start to generate relevant files and folders of the website.

52

Q. Sheng Table 5. Directory information

Directory name

Program language knowledge

Network equipment

Related links

Link file

1.htm

2.htm

1j.htm

Whether there are subordinates

Yes

Yes



Number of subdirectories

2

3

0

Inner frame file

1-00.htm

2-00.htm

Folder

aa

bb

Catalog number

1

2

3

4.3 Analysis of Test Results Using the above test methods and steps, the comparison curve of students’ learning efficiency is obtained, as shown in Fig. 7.

Fig. 7. Comparison curve of student learning efficiency

From the experimental results, it can be seen that when using the cloud computingbased programming language learning website generation system to provide users with learning, the learning efficiency of students is low, and only the learning efficiency of the last test reached nearly 50%; When the programming language learning website generation system provides users with learning, the student’s learning efficiency is very high, and the student’s maximum student efficiency has reached 95%. Therefore, a

Based on the Big Data Program Language Learning Website

53

programming language learning website generation system based on big data can be obtained to improve the learning efficiency of students.

5 Conclusion In view of the low efficiency of students’ learning in the traditional website generation system of program language learning, this paper proposes and designs a website generation system of program language learning based on big data. Under the background of big data, based on the hardware design and software design of the system, the system design of the system is completed, which realizes the generation of the website. The test results show that compared with the cloud based system, the learning efficiency of students is higher, and it has a wide range of application prospects. However, this study only considers the efficiency of students’ learning, and ignores the stability of the system. In the following research, it needs to be improved continuously to make the performance of the website generation system of program language learning reach a higher level.

References 1. Shengran, Z., Changliang, W.: Design and implementation of intelligent unified self-learning system based on SVM. Softw. Guide 8, 127–130 (2019) 2. Lihua, Z.: Design for big data fingerprint recognition system based on ARM and deep learning. J. Hunan Univ. Sci. Technol. (Nat. Sci. Ed.) 34(01), 77–84 (2019) 3. Meng, M., Yinhui, A.: Design of virtual training system for switch machine. Mech. Eng. Autom. 5, 4–6 (2017) 4. Yuyu, Z., Jiaming, L., Xinfei, W.: Design and implementation of resume automatic generation system based on web. Comput. Knowl. Technol. 13(34), 89–92 (2017) 5. Fengliang, H., Ming, X., Wenting, W.: Design on the system of automatic generating experimental report. Exp. Sci. Technol. 9(01), 48–52 (2011) 6. Jinjin, Z., Haihong, E.: Research and implementation of the web page generation system based on responsive web design. Comput. Eng. Softw. 36(06), 37–41 (2015) 7. Huijin, L., Sheng, H., Lihong, W.: Intelligent micro-lecture: the automatic generation system of micro-lecture based on artificial intelligence. Mod. Educ. Technol. 28(11), 5–11 (2018) 8. Wuping, X., Jinyun, X.: Research on generation system from Radl Algorithm to apla program. J. Comput. Res. Dev. 51(04), 856–864 (2014) 9. Baotan, L., Junhong, Q., Haiting, Z.: Style based graph auxiliary generation system in substations. Autom. Electr. Power Syst. 39(14), 120–125 (2015) 10. Lei, X.: Design of translation automatic generation system for English-Chinese machine translation. Mod. Electron. Tech. 41(24), 86–89 (2018)

Evaluation Model of Aquaculture Robot Technology Research Project Based on Machine Learning Peng Cheng(B) Chongqing Vocational College of Transportation, Chongqing 402247, China [email protected]

Abstract. With the increase of research projects of aquaculture robot technology, how to evaluate the research projects of aquaculture robot technology effectively has become the primary task in the research process. However, in the use of the original evaluation model, the problem of improper selection of the index range often occurred. Therefore, the evaluation model of aquaculture robot technology research project based on machine learning is designed. Obtain the evaluation index of aquaculture robot technology project, calculate the index weight to build the evaluation index system, use machine learning algorithm to complete the collection of evaluation samples, and use the above settings to build the evaluation model of aquaculture robot technology research project. At this point, the evaluation model of the aquaculture robot technology research project based on machine learning has been designed. In order to verify the effect of the design model in this article, design a comparison experiment, evaluate the design model and the original model on a project. It can be seen through comparative experiments, it is known from experimental comparison that this model is better than the original model. Keywords: Evaluation of scientific research projects · Index system · Machine learning · Keep improve

1 Preface It is of great significance to improve the efficiency and quality of aquaculture industry to transform the traditional aquaculture industry with modern technology and promote the deep integration of industrialization and informatization into the aquaculture field. Due to the complexity of underwater production environment, research on the application of underwater robots in aquaculture has been paid more and more attention. As a tool for human beings to explore the ocean, underwater robot is a special application of advanced robotic technology in underwater, and is a cutting-edge technology field combining machinery, control, information, navigation, ship and other disciplines [1, 2]. In aquaculture, underwater fishing operation is one of the functional applications of underwater robot. It relies on multi-functional manipulator or straw to catch aquatic © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 54–64, 2020. https://doi.org/10.1007/978-3-030-63955-6_5

Evaluation Model of Aquaculture Robot Technology Research Project

55

products, which can replace artificial operation in dangerous environment. In order to provide a theoretical basis and comprehensive reference for the development of aquatic robot software, and to have a comprehensive understanding of the scientific research project of aquatic robot technology, this research project was evaluated. The evaluation of scientific research projects is an important part of scientific and technological management, and its purpose is to achieve the optimal allocation of scientific and technological resources, improve the investment performance of scientific research projects, and make limited funds play a greater role [3]. With the continuous deepening of the reform of the management system of science and technology plans, the role of performance evaluation of scientific research projects has become increasingly important. From the perspective of national science and technology macro policies, the correct evaluation of scientific research project performance can provide the basis for the state to support various disciplines in basic research and provide important references for supporting decision-making in certain fields of high and new technology. From the perspective of local government science and technology guidance, scientific research project performance Evaluation can provide a basis for local governments to support the transformation of high-tech projects or scientific and technological projects suitable for local economic development, so as to strengthen the management of scientific research projects throughout the process and improve the efficiency of scientific and technological plan management. In this research, a corresponding evaluation model will be set to realize the evaluation of aquaculture robot technology research projects.

2 Evaluation Model of Aquaculture Robot Technology Research Project Based on Machine Learning In the past research, the research project evaluation model of aquaculture robot technology is used to evaluate the research project, but the problem of improper selection of index range often occurs in the selection part of the original model. Therefore, in this paper, machine learning algorithm is used to optimize the original model. The specific optimization structure is as follows (Fig. 1).

56

P. Cheng

Fig. 1. Construction model of aquaculture robot technology research project evaluation model based on machine learning

Through the above process, machine learning and evaluation model are combined to improve the stability of the evaluation model. 2.1 Obtaining Evaluation Indicators for Aquaculture Robot Technology Projects Underwater robots, also known as unmanned underwater vehicles, can perform certain tasks underwater instead of humans. According to the different communication methods with the surface support system, underwater robots can be divided into two categories: remote-controlled underwater robots and autonomous underwater robots [4, 5]. ROV receives remote control commands and power supply from the surface platform via the “umbilical cord” cable; AUV has power energy and intelligent control system, and it can efficiently complete the predetermined tasks by relying on its independent decision-making and control ability [6]. Worldwide applications of underwater robots have expanded to include cable laying and inspection, seabed mineral survey, salvage

Evaluation Model of Aquaculture Robot Technology Research Project

57

operations, underwater archaeology, aquaculture, water environment monitoring and dam inspection of rivers and reservoirs. The primary problem in establishing the model is to transform the contents that need to be evaluated into relevant indicators according to the basic principles and overall requirements of the evaluation and analysis of scientific research projects. The relevant indicators should reflect the main characteristics of the performance evaluation of scientific research projects, and the research process should focus on the research foundation, research level, research effect and other performance evaluation indicators. According to the principle of analytic hierarchy process (ahp), the performance evaluation model of scientific research projects is an organic series composed of several interrelated, complementary, hierarchical and structural comprehensive indicators. The evaluation indexes of the following scientific research projects are obtained by studying the scientific research projects of aquaculture robots, as shown below (Table 1). Table 1. Evaluation index of aquaculture robot technology projectIndex Number Index level 1

Index content

Project management Manpower input

2

Financial input

3

Project preparation

4

Facilities condition

5

Social environment

6 7 8

Project management Project level

Technical index Application of achievements

9

Technical service

10

Monograph

11

Achievements in scientific research

12

Intellectual property right

13

Intellectual property right

14

Project effectiveness Market environment

15

Input output

16

Economic performance

17

Social results

18

Achievement transfer

19

Technology contribution

20

Personnel training

In the process of collecting the above indicators, it is considered that the aquaculture robot is an application project. Therefore, in the evaluation of applied research projects,

58

P. Cheng

more attention should be paid to its scientific value and social value. At the same time, it should closely integrate the needs of economic and social development. It should be guided by technological advancement and driven by market demand. The following standards need to be covered: innovation, practicality, technical theory, key technology, commonality and core high and new technology, independent intellectual property rights, economic benefits, social benefits, etc. 2.2 Calculation of Index Weight and Construction of Evaluation Index System In view of the preliminarily constructed evaluation index system, Delphi screening method was adopted to invite experts for several rounds of opinions consultation, and finally the opinions of experts were summarized to obtain the filtered index system [7]. With built in front of the whole process in scientific projects of college and university evaluation index system is relatively large, the space is limited, so this article only discuss the index system of project phases, in the next chapter in the empirical analysis also make this arrangement, at the same time, the selection of cases, in considering the empirical analysis in this section discuss the index system for the commercial development projects. Set the evaluation index of the past scientific research projects above as the selection range of this evaluation index, and construct a corresponding evaluation index system. Set the index system to P, the evaluation weight to Xi , and the index value to Yi . Then there are: P=

n 

Xa ∗ Ya (a = 1, 2, 3, . . . , m)

(1)

a=1

Among them, a is set as the sequence number of the indicator. Use the above formula to complete the construction of the evaluation index system. This evaluation index system is used as the basis for index selection and weight calculation. In this design, the correlation method is used to assign values to the index weights in the evaluation model, and to set the index system constructed above. Some indicators in the system are set as evaluation influencing factors, and other parts are set as evaluation indicators. Set the evaluation index set to G, and the corresponding reference index set to H , and  G = g1 , g2 , g3 , . . . , gm , H = {h1 , h2 , h3 , . . . , hm }. The types of scientific research projects based on aquaculture robot technology are more complicated, in order to improve the evaluation accuracy of the evaluation model. For the calculation of indicator weights, a dimensionless process is required. The specific formula is as follows: ⎧ g(n)  ⎪ ⎪ g (n) = ⎪ ⎪ j ⎪ ⎪ ⎪ g(m)/n ⎪ ⎨ m=1 (2) h(n) ⎪  ⎪ ⎪ h (n) = ⎪ ⎪ j ⎪ ⎪ ⎪ h(m)/n ⎩ m=1

Evaluation Model of Aquaculture Robot Technology Research Project

59

In the formula: n represents the number of indicators in the indicator system. m represents the index number, and the above formula is used to calculate the index weight in the system. And reorder the metrics. Due to the large range of indicators, only the weights of indicator types are sorted in this section. The specific results are shown below (Table 2). Table 2. Index weight setting Index weight sorting Index category

Index specific gravity

1

Project effectiveness 35%

2

Project management 35%

3

Project level

30%

Using the above evaluation index weights, complete the evaluation model construction of aquaculture robot technology research projects based on machine learning. 2.3 Using Machine Learning Algorithms to Complete Evaluation Sample Collection The relevant information in the collected project is used as the evaluation data sample. The feature vector [8] of the known sample is c = {c1 , c2 , c3 , . . . , cn } and the category label is d = {d1 , d2 , d3 , . . . , dn }. According to the Bayesian formula, the conditional probability (post-test probability) that the sample belongs to each class is: T (c|d ) =

T (c|d )T (c) T (d )

(3)

For all classification labels d1 , d2 , d3 , . . . , dn , T (d ) is the same. When comparing T (c|d ) and T (cn |d ), you can ignore the T (d ) parameter. So the discriminant function of the classifier is: arg maxT (cn |d )T (d )

(4)

Based on the above data, a given data sample is processed, and the sample content is measured to obtain the corresponding sample processing result. Assuming that the sample obeys the Gaussian distribution, the parameters of the prior probability distribution are determined during training, usually using the maximum likelihood estimation, that is, maximizing the log-likelihood function. To improve the processing effect of sample data, a simple support vector machine linear classifier was created. The optimization goal of the support vector machine is to find a line so that the nearest point to the line can be the farthest. In addition, the points marked in blue are the key support points for fitting the data and are called support vectors [9, 10]. The key factor that the classifier can successfully fit is the position of these support vectors. Data points far from the boundary have no effect on the classifier. Through this section, the classification of scientific research project data and information is achieved.

60

P. Cheng

2.4 Construction of Evaluation Model for Scientific Research Project of Aquaculture Robot Technology The data processing and index processing of the rating model are completed through the above sections. Using the data obtained above, the construction of an aquaculture robot technology research project evaluation model is completed. The specific process is shown below (Fig. 2).

Fig. 2. Evaluation model building process

After calculating the weight of the evaluation index, using the method of weighted function to calculate the index score value of the aquaculture robot technology research project, combined with the evaluation grade table, to complete the construction of the aquaculture robot technology research project. Considering the counting habit of the percentage system, multiply the weight of each index by 100 to become the standardized value of the evaluation result. The final score of each sub index is: An =

l 

Eom χomi

(5)

i=1

In the formula: Eom represents the o-th weight value of the m-th sub-indicator, and χomi represents the score value of the i-th index by the o-th expert. An represents the quantified score of the calculated i-th index. The final score of the comprehensive indicator is: l

A=

i=1

i

Ai (6)

Evaluation Model of Aquaculture Robot Technology Research Project

61

In the formula: A is the final score. The evaluation grade is established according to the expert score. Combining the above formula with index weight, the corresponding evaluation results are obtained. The results are as follows. Table 3. Classification of evaluation results Score

Scale Grade

[80,100] I

Good

[60,80]

II

Preferably

[40,60]

III

Commonly

[20,40]

IV

Poor

[0,20]

V

Difference

Take Table 3 as the standard and evaluate the pros and cons of the aquaculture robot technology research project according to the score value calculated by Formula 6. At this point, the evaluation model of the aquaculture robot technology research project based on machine learning has been designed.

3 Test Experiment In the above part, the design of the evaluation model of the aquaculture robot technology research project based on machine learning is completed. In order to ensure that the evaluation model designed has high-precision evaluation results, the evaluation effect is tested in the form of experiments. 3.1 Experimental Content In the process of this experiment, an aquaculture robot research project was selected as the object of this experiment, and the project was evaluated using the design evaluation model and the original evaluation model, and the accuracy of the evaluation results was compared. The selection range of known evaluation indexes directly affects the accuracy of evaluation results. In this experiment, the index selection range of the two evaluation models will be compared to reflect the evaluation accuracy of the model. During the experiment, part of the calculation process was involved. In order to improve the calculation capacity in the experiment, the equipment used in the set experiment is shown below (Fig. 3).

62

P. Cheng

Fig. 3. Experimental equipment

Use the above experimental equipment to complete the experimental process. In the process of this experiment, the corresponding experimental samples will be set. Through the comparison results between the evaluation index range of the evaluation model and the experimental samples, the differences between the design model and the original model will be obtained. 3.2 Experimental Samples The design of the experimental process is completed through the above part. In order to improve the reliability of the experimental results, the experimental samples are set as follows (Table 4). Table 4. Experimental samples Experiment sample No

Sample type

Sample content

1

Project management

Manpower input

2 3

Financial input Project level

4 5

Technical index Application of achievements

Project effectiveness

Achievements in scientific research

6

Intellectual property right

7

Achievement transfer

8

Technology contribution

The above experimental sample was used as the control group of this experiment, and the above experimental equipment was used to complete the experimental process and obtain the experimental results. The results of this experiment are presented in the form of a table.

Evaluation Model of Aquaculture Robot Technology Research Project

63

3.3 Experimental Results With the above settings, the experimental process is completed. The specific experimental results are shown below (Table 5). Table 5. Experimental results Sample content

Does the design model include this index

Whether this indicator is included in the original model

Manpower input

Y

Y

Financial input

Y

N

Technical index

Y

N

Application of achievements

Y

N

Achievements in scientific research

Y

Y

Intellectual property right

Y

Y

Achievement transfer

Y

Y

Technology contribution

Y

Y

According to the above experimental results, the index range of the design model in this paper covers the index of experimental samples. The index coverage of the original model does not fully cover the index of experimental samples. Therefore, the index range of the designed evaluation model is larger than that of the original evaluation model. The evaluation precision of the model can be improved by designing this part. The index range of the designed model is of great significance to the evaluation results of scientific research projects.

4 Concluding Remarks The development of social economy and the need of the development of scientific research projects put forward a strong demand for the evaluation of scientific research activities. Scientific research project is the main form of R & D activities, and project evaluation is the most basic and main form of scientific research evaluation. Compared with general projects, scientific research projects have many special features, such as the difficulty and asymmetry of information collection, the difficulty of process control, and the inconspicuous form of output, which increase the difficulty of evaluation. In this study, machine learning algorithm is used to reduce the difficulty of research on aquaculture robot technology and improve the reliability of evaluation.

References 1. Xiaoyu, M., Xiuhua, J.: Analysis for the generalization ability of machine learning based full reference image quality metrics. J. Commun. Univ. China Sci. Technol. 26(04), 42–49 (2019)

64

P. Cheng

2. Cuixia, F., Yiyong, L.: Universities theoretical and experimental teaching quality evaluation model based on RVM machine learning method. Mod. Electron. Tech. 42(13), 181–186 (2019) 3. Xianghua, F., Xiaoyu, S., Zhilong, C.: Evaluation of Chunhua county’s water resources carrying capacity prediction. J. Xi’an Univ. Tech. 34(04), 447–453 (2018) 4. Qi, Z., Mingyu, L.: Regional ecological vulnerability assessment based on VSD Model—the case of Yanbian Korean Autonomous Prefecture. J. Agric. Sci. Yanbian Univ. 40(04), 7–15 (2018) 5. Min, D., Debao, W., Suya, Z.: Risk assessment of highway slope geological disasters based on GIS. J. Changsha Univ. Sci. Technol. (Nat. Sci.) 15(04), 59–65 (2018) 6. Bangyang, W., Jing, L., Jian, S.: Evaluation on the efficiency of transit network operation based on analytic hierarchy process and multivariate nonlinear function. J. Hefei Univ. Technol. (Nat. Sci.) 41(12), 1684–1689 (2018) 7. Ximing, P.: On reform of teachers’ scientific research performance assessment in higher vocational colleges. J. Wuhan Inst. Shipbuilding Technol. 17(04), 76–79 (2018) 8. Jinlong, Z., Weiwei, Z., Fuming, S.: Analysis and evaluation of research efficiency of university collaborative innovation center based on PCA-SEDEA: take the collaborative innovation center of industry and industry in Jiangsu as an example. Sci. Technol. Manag. Res. 38(24), 65–72 (2018) 9. Zhongtai, Y.: Study on homogenization of scientific research activities in colleges. J. Baoji Univ. Arts Sci. (Soc. Sci. Ed.) 38(06), 108–113 (2018) 10. Min, Z.: Evaluation model of flight delays Early warning set pairing based on game theory weight. Comput. Simul. 36(10), 88–92+188 (2019)

Research on the Application of Artificial Intelligence Technology in the Quality Evaluation of Experiential Physical Education Tian Xiong(B) Physical Education Department of Shaanxi, University of Traditional Chinese Medicine, Xianyang 712046, China [email protected]

Abstract. In order to improve the teaching quality of experiential physical education, a new method of teaching quality evaluation is designed, and the quantitative analysis of quality evaluation is realized by using artificial intelligence technology. First determine the teaching quality evaluation standard. In the process of physical education teaching, we use artificial intelligence technology to collect motion data, and realize the recognition and analysis of motion data from two aspects of image and voice. The evaluation system of experiential physical education quality is constructed and the evaluation index is set under the system. The specific value of the evaluation index and the corresponding weight value are calculated respectively, the quantitative result of the comprehensive evaluation index is compared with the evaluation standard, and finally the evaluation result of the quality of physical education is obtained. Through the test and analysis of the applied experiment, it is found that the application of artificial intelligence technology to the teaching quality evaluation method can effectively improve the authority of the evaluation results and indirectly improve the teaching quality of college physical education. Keywords: Artificial intelligence technology · Experiential teaching · Physical education quality · Teaching quality evaluation

1 Introduction Sports is a complex social and cultural phenomenon. It is a conscious, purposeful and organized social activity that promotes comprehensive development, improves life style, and improves physical quality and sports ability according to the laws of human body growth and development, skill formation and function improvement [1]. Physical education can be divided into mass sports, professional sports, school sports and other types, in which school sports is mainly through the teaching of teachers, so that students can master the relevant sports skills, and improve the physical quality of students. In order to improve the teaching quality of physical education in schools, the teaching methods of physical education curriculum have been changed gradually. Experiential education is a new form of training and education. It is based on the goal of moral education © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 65–76, 2020. https://doi.org/10.1007/978-3-030-63955-6_6

66

T. Xiong

and the psychological and physiological characteristics of minors as well as their personal experiences. Let the juveniles experience and feel in real life, and form their own moral consciousness and ideological quality through reflection experience and experience internalization, and accumulate their own ideological and moral behavior through repeated experience. Minors dominate themselves and correct themselves in various experiences. They experience, feel and construct the patriotic feelings, national spirit and collective consciousness that the society and The Times expect them to have in their daily behaviors and interactions with others. Experiential physical education can be divided into four stages: personal experience, formation, test and reflection, which can encourage students to produce new experience and new understanding, and thus develop their ability to adapt to nature and society and form a positive attitude towards life. In order to achieve good teaching results of experiential physical education, teachers’ teaching quality should be evaluated regularly. The evaluation of physical education teaching quality is a very important part of the current reform of physical education teaching. Many experts and scholars have made fruitful research in this area, which has promoted the development of physical education evaluation theory and practice [2]. But at the same time, we should realize clearly that with the continuous deepening of the reform of physical education in colleges and universities, the current teaching quality evaluation system can not meet the needs of the current educational situation, and there are still many problems, which need to be further studied and improved. At present, the main technical methods used in the quality evaluation of experiential physical education include literature, questionnaire, analytic hierarchy process and so on. Due to the use of these techniques, the relevant factors of physical education evaluation are not clearly defined, the evaluation dimension is not comprehensive, and the evaluation index validity is low. Therefore, the evaluation results obtained by the traditional evaluation methods have no reference value. In order to solve the above problems in traditional methods, artificial intelligence technology is introduced. Artificial intelligence is the study of the computer to simulate some of the thinking process and intelligent behavior of the subject, mainly including the computer to realize the principle of intelligence, the manufacture of computers similar to human brain intelligence, so that the computer can achieve a higher level of application. Artificial intelligence will involve computer science, psychology, philosophy and linguistics. Applying this artificial intelligence technology to the evaluation method of experiential physical education quality can realize the quantitative analysis of evaluation indexes and obtain more intuitive evaluation results. Through the analysis of the current physical education quality evaluation system, find out its existing problems, use artificial intelligence technology to solve the existing problems, in order to provide theoretical and practical reference for the future experiential physical education reform and physical education quality evaluation.

Research on the Application of Artificial Intelligence Technology

67

2 Design of Quality Evaluation Method for Experiential Physical Education Teaching The evaluation of physical education itself is a complex system, which is composed of a large number of subsystems and elements. Physical education evaluation is a process, which is accompanied by the transmission of teaching information, evaluation information and feedback of evaluation information. Therefore, the evaluation of physical education is regarded as a system, and the problems in the evaluation of physical education are analyzed from the perspective of systematic scientific theory [3]. Therefore, in the actual process of experiential PE teaching quality evaluation, the system theory, information theory, cybernetics and the whole principle, order principle and feedback principle in the system science theory are the theoretical basis for the research and analysis of this article. Based on the above principles, artificial intelligence technology is introduced to optimize the design of experiential PE teaching evaluation method. The optimized teaching quality evaluation process is shown in Fig. 1. Teaching quality evaluation of unified arrangement of teaching courses Determine the evaluation standard of physical education teaching quality Collection of teaching data by artificial intelligence technology Analysis and collection of data by artificial intelligence technology Set up the evaluation index of Physical Education Teaching Teaching evaluation index and weight calculation By comparing the calculation results with the standards, the teaching quality evaluation results are obtained Fig. 1. Flow chart of experiential physical education quality evaluation

According to the quality evaluation process shown in Fig. 1, the final teaching quality evaluation results are obtained by comprehensive consideration of teaching attitude, teaching ability and teaching effect. 2.1 Determine the Evaluation Criteria for Teaching Quality The teaching attitude, ability and effect of experiential PE teachers are integrated, and the indicators reflecting teaching quality are quantified through the established teaching

68

T. Xiong

quality evaluation index system. Finally, the evaluation results of teaching quality are converted into scores and divided into different teaching quality levels according to scores [4]. The evaluation standards for the quality of experiential physical education are shown in Table 1. Table 1. Evaluation standards of physical education teaching quality Classification

Excellent Good

Score

≥85

Qualified Unqualified

76–84 60–75

θ

(5)

The fuzzy C-means clustering method is adopted to construct a feature extraction model of complex educational resource information distribution network data, and data fusion processing is carried out according to the feature extraction results. A four-tuple structure is used to describe   the information correlation characteristics of educational resource information: FP Xij , Pij , (supk1 (D), · · · , supkf (D)), (Tk1 , · · · Tkj ) , wherein Xij is the correlation dimension of the characteristic information flow of educational resource information at the moment, Tij is the mutual information amount [12]. By

Encryption and Compression Storage Method

81

the spatial regional information fusion method, the optimal probability of educational resource information distribution is obtained, which is the fuzzy feature set of educational resource information. By spectral analysis method, the multidimensional reconstruction output of the obtained data is as follows: i xO = xSi + Kdimax (xLi − xSi ) (6)  i  Wherein: K = 1/xL − xSi , according to the association rule set of educational resource information, for the calculation, the similarity feature of educational resource information is extracted, and combined with semantic information clustering method, the high-resolution feature reconstruction output of educational resource information is obtained as follows: ⎧ ⎫ l1 l2 ⎨ ⎬ f (x) = sgn z αi+ yi K(xi , x) + αi− yi K(xi , x) + b (7) ⎩ ⎭ i=1

i=1

In the formula, αi+ represents the high-dimensional distribution feature quantity of educational resource information, αi− is the low-frequency component of educational resource information, and l1 and l1 respectively represent the coupling feature quantity of relevant information. Combined with the fuzzy feature mining method, the gradient vector of network data is obtained as follows: 1 |GX (x, y)| m×n n

AVGX =

m

(8)

x=1 y=1

Where, m and n are the geometric feature similarity of educational resource information respectively, D is set as the spatial distribution domain, and T1 is the sampling time delay of educational resource information. According to the above analysis, a feature extraction model of educational resource information is established, and encrypted, compressed and stored according to the distributed detection results of educational resource information features [13].

3 Optimization of Encryption Compression Storage Model for Complex Educational Resource Information Distribution Network 3.1 Compression of Dynamic Data of Educational Resources On the basis of the above-mentioned construction of the dynamic feature information collection and fuzzy clustering model of educational resource information, and the sampling and feature extraction of educational resource information using the wireless sensor array of the complex network, data dynamic compression is carried out [14], and a educational resource encryption compression storage model based on the complex network and large data analysis is proposed. The output data packets of the feature compression are: sgn(zR2 (k) − RMDMMA_R ) = sgn(zR2 (k) − eˆ R2 (k))

(9)

82

J. Ma et al.

sgn(zI2 (k) − RMDMMA_I ) = sgn(zI2 (k) − eˆ I2 (k))

(10)

Wherein, eˆ R2 (k) represents the observation sequence of ocean resource information flow, zR2 (k) is the signal-to-noise ratio in the original training set, zI2 (k) is the covariance function, eˆ I2 (k) is the discrete distribution reconstruction output of ocean resource information, feature combination and automatic information extraction are carried out on collected ocean resource information, metadata structure features of ocean resource information are extracted, probability density statistics of ocean resource information are carried out by combining metadata analysis methods, and information components of dynamic data compression output are obtained as follows: ηcomm =

 k1 · l  · 1 − pdrop Ecomm

(11)

Wherein, pdrop is an association rule distribution function of educational resource information, the output power loss is CHi (i ∈ C1 ), the characteristic component of the information difference degree of the educational resource information is extracted, the optimized dynamic compression reconstruction of the educational resource information is realized, and the output is as follows: ⎧ ⎫ l ⎨ ⎬ f (x) = sgn αj∗ yj K(x, xj ) + b∗ , x ∈ Rn (12) ⎩ ⎭ j=1

Wherein, b∗ = yi −

l  j=1

 yj αj K(xj , xi ), i ∈ {i0 < αi∗ < u(xi )C} . According to

the above method, the dynamic structural feature values of ocean resource information are extracted, and the data structure is reorganized through dynamic data compression and information perception [15]. 3.2 Encrypted and Compressed Storage of Educational Resources Using the distributed fusion method of multi-dimensional feature space structure, the educational resources are encrypted and compressed and the information coding is adaptively distributed, and the fuzzy clustering distribution model of data is obtained as follows: ⎫ l ⎪ ⎪ 1 ⎪ u(xj )ξj ⎪ min w2 + C ⎪ ⎬ w,b,ξ 2 j=1 (13) ⎪ s.t. yj ((w · xj ) + b) + ξj ≥ 1⎪ ⎪ ⎪ ⎪ ⎭ ξj ≥ 0, j = 1, 2, . . . , l According to the mutual coupling relationship of educational resource information, high-dimensional phase space reconstruction is introduced, feature combination and automatic information extraction are carried out on the collected educational resource

Encryption and Compression Storage Method

83

information, multi-resolution feature quantities of the educational resource information are extracted, and statistical information output by educational resource encryption compression and information encoding is obtained as follows: Ghi 1 ∂ui =  ( − βc1 ) 2 ∂pi hj pj + σ 1 + γi

(14)

j=i

The kernel functions of educational resources encryption and compression and information encoding are obtained by using the least square programming algorithm. The optimized programming functions of encryption and compression storage are as follows: ⎫ l l l ⎪ 1 ⎪ ⎪ min yi yj αi αj K(xi , xj ) − αj ⎪ ⎪ α 2 ⎪ ⎪ i=1 j=1 j=1 ⎪ ⎬ l (15) ⎪ ⎪ s.t. yj αj = 0 ⎪ ⎪ ⎪ ⎪ j=1 ⎪ ⎪ ⎭ j = 1, 2, . . . , l 0 ≤ αj ≤ u(xj )C Based on fuzzy membership analysis, the output detection probability of dynamic compression and reconstruction of educational resource information is obtained as follows: Pd =

K

ck e−j2π fkT



(16)

k=−K

Wherein, ck represents the multi-information feature distribution set of educational resource information. Then the simplified mathematical model of educational resource encryption and compression and information encoding can be described by the following formula: ⎧ G1 = b11 a1 + b12 a2 + . . . + b1n an ⎪ ⎪ ⎪ ⎨ G2 = b21 a1 + b22 a2 + . . . + b2n an (17) ⎪ ... ... ... ⎪ ⎪ ⎩ Gn = bn1 a1 + bn2 a2 + . . . + bnm an Wherein, Gj and Gk have strong correlation, Gj represents the oscillation value and interference term of educational resource information, Gk is the principal component of educational resource information. To sum up, the algorithm is designed to optimize the encryption and compression storage model of educational resources.

4 Simulation Experiment and Result Analysis In order to verify the application performance of the method in realizing the encryption and compression storage of the complex educational resource information distribution

84

J. Ma et al.

network, simulation test analysis is carried out, a phase space model of the encryption and compression storage of the educational resource is established, data analysis is carried out by Matlab, assuming that the sampling length of the educational resource information is 1024, the size of the data is 200, the size of the test set is 120, and the attribute category number of the data of the complex educational resource information distribution network is 65. For the initial frequency f1 = 1.52 Hz of data and the ending sampling frequency f2 = 2.43 Hz, educational resources are encrypted, compressed and stored according to the above parameter settings to obtain a educational resources information test set as shown in Fig. 2. 4 3 2

Amplitude

1 0 -1 -2 -3 -4 -5

0

100

200

300

400 Time/s

500

600

700

800

Fig. 2. Educational resources information test set

Taking the data in Fig. 2 as the research object, the educational resources are encrypted, compressed and stored, and the educational resources are encrypted, compressed and adaptively allocated for information coding by adopting the distributed fusion method of multi-dimensional feature space structure. The dynamic reconstruction output is shown in Fig. 3. Analysis of Fig. 3 shows that this method can effectively realize the encryption and compression storage of educational resources, and the dynamic compression performance of data is better. after testing the precision ratio and recall ratio of different methods for encryption and compression storage of educational resources, it is concluded that the precision ratio and recall ratio of this method are increased by 15.3% and 21.8%, respectively. the characteristic compression capability of encryption and compression storage of educational resources using this method is better, the precision of encryption and transmission of resource information is higher, and the precision ratio and recall ratio of educational resource information are improved. Test time cost, the comparison results are shown in Table 1. Analysis of Table 1 shows that the time cost of educational resource encryption and compression storage in this method is relatively short.

Encryption and Compression Storage Method

85

1

z

0.5

0

-0.5

-1 1 1

0.5 0.5 0 y

0 -0.5

-0.5

x

Fig. 3. Encrypted compressed storage output

Table 1. Comparison of time expenses (unit: ms) Data size/Gbit

This method

Reference [3]

Reference [6]

10

13.5

34.7

54.5

20

15.6

83.4

64.5

30

23.5

96.7

83.4

40

27.8

110.3

95.6

5 Conclusions In this paper, the data detection optimization design of the complex educational resource information distribution network is carried out, and the information reconstruction of the complex educational resource information distribution network is carried out in combination with the compression processing technology of the educational resource information, so as to improve the mining capability of the educational resource information. In this paper, the educational resource encryption compression storage model based on the complex network and large data analysis is proposed. Combined with spatial region reconstruction method, statistical analysis of complex educational resource information distribution network data is carried out, multi-dimensional random linear coding technology is adopted, information fusion of complex educational resource information distribution network data is carried out, fuzzy C-means clustering method is adopted, feature extraction model of complex educational resource information distribution network data is constructed, high-dimensional phase space reconstruction is introduced, feature combination and automatic information extraction are carried out on collected educational resource information, multi-resolution feature quantity of educational resource information is extracted, and educational resource encryption compression and adaptive

86

J. Ma et al.

allocation and reconstruction of information coding are carried out. The analysis shows that the method of this paper has higher precision and better feature resolution, improves the precision and recall of data, and reduces the processing overhead.

References 1. Colmers-Gray, I.N., Krishnan, K., Chan, T.M., et al.: The revised METRIQ score: a quality evaluation tool for online educational resources. AEM Educ. Train. 38(6), 1591–1595 (2019) 2. Liu, T., Zheng, L., Du, P.: Spatial characteristics and distribution pattern of the equilibrium level of municipal compulsory education resources—take dalian primary school as an example. Econ. Geogr. 19(1), 95–131 (2018) 3. Shaikh, N., De Azevedo, R.U., Rajesh, A., et al.: Re: lack of online video educational resources for open colorectal surgery training. ANZ J. Surg. 89(5), 618 (2019) 4. Nathan, J., Bound, C., Watson, M.: G41(P) The use of peer design in the development of educational resources for adolescent type 1 diabetes patients. Arch. Dis. Child. 104(2), A17– A22 (2019) 5. Xiao, H.: MR-based synthetic CT generation using a deep convolutional neural network method. Med. Phys. 44(4), 1408–1419 (2017) 6. Tanious, A., Brooks, J.D., Wang, L.J., et al.: Educational resources for vascular laboratory education in vascular surgery residencies and fellowships: survey of vascular surgery program directors. J. Vasc. Surg. 15(25), 25–31 (2019) 7. Fakhar, M., Mahyarinia, M.R., Zafarani, J.: On nonsmooth robust multiobjective optimization under generalized convexity with applications to portfolio optimization. Eur. J. Oper. Res. 265(1), 39–48 (2018) 8. Sun, X.K., Li, X.B., Long, X.J., et al.: On robust approximate optimal solutions for uncertain convex optimization and applications to multi-objective optimization. Pac. J. Optim. 13(4), 621–643 (2017) 9. Jung, I., Lee, J.: A cross-ultural approach to the adoption of open educational resources in higher education. Br. J. Edu. Technol. 23(1), 156–175 (2019) 10. Stovall, J.P., Laird, S.G., Lana, W., et al.: Student and instructor generated open educational resources compare favorably to a traditional textbook. J. Forest. 4(4), 4–10 (2019) 11. Monroe, K.S., Evans, M.A., Mukkamala, S.G., et al.: Moving anesthesiology educational resources to the point of care: Experience with a pediatric anesthesia mobile app [J]. Korean J. Anesthesiol. 71(3), 25–34 (2018) 12. Li, L., Cheng, S., Wen, Z., et al.: A debris flow prediction model based on the improved KPCA and mixed kernel function LSSVR. Inf. Control 48(5), 536–544 (2019) 13. Hilton, J.: Open educational resources, student efficacy, and user perceptions: a synthesis of research published between 2015 and 2018. Educ. Tech. Res. Dev. 12(2), 124–129 (2019) 14. Darwish, H.: Open educational resources (OER) Edupreneurship business models for different stakeholders. Educ. Inf. Technol. 15(1), 1–32 (2019) 15. Shaikh, N., et al.: Re: lack of online video educational resources for open colorectal surgery training. ANZ J. Surg. 12(3), 145–150 (2019)

High-Quality Extraction Method of Education Resources Based on Block Chain Trusted Big Data Hao Zhang1(B) , Bin Zhao2 , and Ji-shun Ma3 1 School of Economics, Zhejiang University, Zhoushan 316000, China

[email protected] 2 Department of Electronic Science and Technology,

Huazhong University of Science and Technology, Zhuji 311800, China 3 Pharmaceutical Engineering, Zhejiang University of Chinese Medicine, Zhuji 311800, China

Abstract. In order to improve the level of educational administration, it is necessary to extract educational resources with high quality Based on the block chain trust, a high-quality education resource extraction method is proposed, and a model function of high-quality education resource extraction is designed by using the spatial distribution resource scheduling model. In order to judge the convergence of high-quality education resource extraction process, a statistical analysis model of high-quality education resource extraction is established. The method of quantitative feature analysis and fuzzy information clustering is used to extract and control the high quality of educational resources. The root game equilibrium optimization algorithm realizes the optimization of the high quality of educational resources. The simulation results show that the optimization ability of using this method to extract the high quality of educational resources is better, and the scheduling process has strong convergence, it improves the ability of optimal scheduling and acquisition of educational resources. Keywords: Block chain · Trusted big data · Administrative resources · High quality extraction · Data access

1 Introduction Since the start of IT informatization construction in education departments, education data have shown explosive growth, accumulated for so many years, formed a large amount of data already, and naturally formed “education big data”. The core goal of data governance is to sort out and standardize education data, solve possible problems, and establish unified, real-time and accurate authoritative data resources for all departments and functional units to share. This paper studies the high-quality extraction model of education resources, brings the national advantageous education resources into the dispatching scope, and combines the information fusion and big data analysis technology to carry out the optimal design of the high-quality extraction scheme of education © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 87–96, 2020. https://doi.org/10.1007/978-3-030-63955-6_8

88

H. Zhang et al.

resources [2]. People pay attention to the optimal dispatching and feature extraction design of related education resources. In the high-quality extraction and scheme design of education resources, the dispatching model design is the key. The statistical analysis model of high-quality extraction of education resources is established. Combined with fuzzy information fusion and optimization method, the operational research analysis theory [1] is adopted to carry out the high-quality extraction and game planning design of education resources [3]. This paper proposes a high-quality extraction scheme of education resources based on block chain trusted big data. Firstly, a spatial distributed resource scheduling model is adopted to design a mathematical model for high-quality extraction of education resources. Secondly, an optimal optimization parameter design for high-quality extraction of education resources is carried out. Combining with the concentration of block chain trusted big data distribution information, the fitness level and equilibrium level of high-quality extraction of education resources are analyzed. The root game equilibrium optimization algorithm realizes the optimization of high-quality extraction of education resources. Finally, the simulation test analysis shows the superior performance of the method in improving the high-quality extraction ability of education resources [4].

2 Mathematical Model and Parameter Analysis for High Quality Extraction of Education Resources 2.1 High Quality Extraction Model of Education Resources In order to realize the scheduling optimization of education resources, a mathematical model design for high-quality extraction of education resources is carried out in combination with an adaptive optimization method [5]. A quadratic programming model for high-quality extraction of education resources is established. The distribution center (i, j = 0) using high-quality extraction of education resources has K (K = 1,…, K) scheduling paths, N(i, j = 1,…, N) scheduling points, and the initial load capacity of each group of education resources is Q.  1 k>j Define the following variables and parameters: xijk = 0 else gi : Required distribution quantity of education resources characteristic distribution point i. ui : Distribution service order of education resources characteristic distribution point I. Using the model correlation variable analysis method, the multi-objective function for high-quality extraction of education resources can be determined as follows: N  K N   ↔ min D = d ij xijk

(1)

i=0 j=0 k=1

min C =

N  K N   i=0 j=0 k=1

cij xijk

(2)

High-Quality Extraction Method of Education Resources

max S =

N  K N  

sij xijk

89

(3)

i=0 j=0 k=1

As shown in the above formula, there are three optimization objectives for the highquality extraction of education resources [6]. By adopting a linear programming scheme, the constraint conditions for obtaining the shortest high-quality extraction distance, the lowest cost and the largest safety factor of education resources are as follows: N 

gi yik ≤ Q, k = 1, 2, . . . , K

(4)

i=0 K  k=1

Where,

N  i=0

 yik =

1 i = 1, 2, . . . , N K i=0

(5)

xijk = yjk j = 0, 1, 2, . . . , N; k = 1, 2, . . . , K K 

xijk = yik i = 0, 1, 2, . . . , N; k = 1, 2, . . . , K

j=0

xijk = 0 ∀i = j; k = 1, 2, . . . , K ui − uj + (N + 1)



(6)

xijk ≤ N

k

i = j, i = 1, 2, . . . , N; j = 1, 2, . . . , N

(7)

Where, xijk ∈ {0, 1}; yik ∈ {0, 1}; ui >0 Under the constraint conditions, the process of education resources distribution is optimized. Ant colony spatial planning algorithm is adopted to obtain the education resources scheduling scheme with K scheduling paths. Combined with the adaptive optimization method, high-quality extraction and spatial planning design of education resources are carried out [7]. 2.2 Parameter Optimization for High Quality Extraction of Education Resources The convergence judgment of the high-quality extraction process of education resources is carried out by combining a linear programming scheme, a statistical analysis model for high-quality extraction of education resources is established, emergency parameter optimization design of education resources is carried out [8–10], the number of ants in ant colonies for high-quality extraction of education resources is set to be m, the spatial planning path distribution constraint parameter from the characteristic distribution point I of education resources to the characteristic distribution point j of education resources ↔

is ηij , d ij is taken, Indicates the shortest path from the feature distribution point I of

90

H. Zhang et al.

education resources to the feature distribution point J of education resources. Combined with the intensity analysis of pheromone concentration, automatic optimization of highquality extraction of education resources is carried out. The spatial planning model for high-quality extraction of education resources is obtained as follows: ⎧ β −γ τij (t)α ηij cij sijσ ⎨ if j ∈ allowedk (t)  α β −γ σ Pijk (t) = (8) j∈allowedk (t) τij (t) ηij cij sij ⎩ 0 else In the above formula, allowedk (t) = (1, 2, . . . , n)−tabuk indicates the convenient point for high-quality extraction of education resources. Ant K is allowed to select the next point. Obtaining a search list tabuk (k = 1, 2,…, m) for high-quality extraction of education resources, and obtaining a diversity space analytic function for high-quality extraction of education resources by recording the path of ant k at time t and combining a fuzzy degree monitoring method as follows:

 β −γ arg maxj=allowedk (t) τij (t)α ηij cij sijσ q ≤ q0 j = (9) J else q0 ∈ (0, 1) is the spatial distribution constant for high-quality extraction of education resources, and q is a random number between 0 and 1. When q > q0 , the optimization threshold j for high-quality extraction of education resources satisfies the convergence condition [11], and the global information update formula for high-quality extraction of education resources is as follows: τij = (1 − ρ)τij + ρ τijbest , ρ ∈ (0, 1)

(10)

τijbest = 1/Obest

(11)

In the formula, Obest is the global optimal target value for current ants to perform high-quality extraction of education resources. ρ is the pheromone concentration. After obtaining the optimal solution for high-quality extraction of education resources by combining ant colony optimization method, in the spatial distribution vector set [12], the distribution center optimization parameters for high-quality extraction of education resources are obtained as follows: (12) τij = (1 − ξ )τij + ξ τ0 , ξ ∈ (0, 1) 

3 3 In the formula, τ0 is constant, τ0 = 1/ nOmin , Omin are global update disturbance thresholds. Based on the above analysis, a parameter optimization model for high-quality extraction of education resources is established to improve the optimal scheduling and access capability of education resources [13].

3 Optimal Design for High Quality Extraction of Education Resources 3.1 Optimization of High Quality Extraction of Education Resources The auto-correlation quantitative feature analysis and fuzzy information clustering method are adopted to carry out high-quality extraction and adaptive optimization control

High-Quality Extraction Method of Education Resources

91

of education resources, a block chain trusted big data detection algorithm is established to realize optimization design of high-quality extraction of education resources [14], game control of high-quality extraction of education resources is carried out according to the distribution relevance of block chain trusted big data, and the game model of high-quality extraction of education resources is as follows: ⎧ α β ⎪ ⎨ [τij (t)] [ηijα(t)] β , if j ∈ allowedk (13) pijk (t) = s∈allowedk [τis (t)] [ηis (t)] ⎪ ⎩ 0, else τij (t + n) = (1 − ρ)τij (t) + τij (t)

(14)

τij (t) = τijk (t)

(15)

 τijk (t)=

Q Lk ,

0,

K >L else

(16)

In the formula, τij (t) and pijk (t) respectively represent the pheromone intensity on the connection line between node I and node j of high-quality extraction of education resources at time t, and the conditional distribution probability of high-quality extraction of education resources represents the relative importance of ant k trajectory; ηij (t) represents the conditional probability of ant K selecting a high-quality extraction path of education resources; ηij (t)= d1ij represents the expected degree of ant K transferring from high-quality extraction node I of education resources to scheduling node J. Usually, taduk represents the spatial distance of emergency network of education resources, and fuzzy degree search method is adopted to obtain the conditional permission function of high-quality extraction of education resources as K, d represents the optimization ability of ant K, taduk represents the pheromone concentration recorded by ant K for optimization, set represents the ant colony variability of high-quality extraction of education resources, and adopts ant colony optimization method to obtain the information amount on path ij in the process of high-quality extraction of education resources. Q indicates the pheromone intensity of education resources emergency, and k is constant. According to the above analysis, an ant colony optimization model for high-quality extraction of education resources is established, and spatial planning and design are carried out according to the ant colony optimization results [15]. 3.2 Implementation Process of High Quality Extraction of Education Resources A block chain trusted big data detection algorithm is established to realize the optimization design of high-quality extraction of education resources, game control of highquality extraction of education resources is carried out according to the distribution relevance of block chain trusted big data, and fitness function is set. Since the problem of high-quality extraction of education resources is transformed into a Traveling Salesman Problems (TSP), according to TSP problem, the optimized planning model is as follows:  xi,j di,j (17) min Dtotal = i∈N j∈N

92

H. Zhang et al.

⎧  xi,j = 1, ∀j ∈ N ⎪ ⎪ ⎪ j∈N ⎪ ⎨  xi,j = 1, ∀i ∈ N s.t. i∈N ⎪   ⎪ ⎪ ⎪ xi,j ≥ 1, M ⊂ N ⎩

(18)

i∈M j∈N −M

In the formula, di,j is the distance between the two network nodes i, j for highquality extraction of resources is expressed, the game control for high-quality extraction of education resources is carried out according to the association of trusted large data distribution in the block chain, and the fitness level and equilibrium level for highquality extraction of education resources are analyzed in combination with the concentration of trusted large data distribution information in the block chain. The algorithm implementation steps are as follows: Step (1) Planning and coding the high-quality extraction of education resources. The linear programming coding path of the commonly used planning path is represented by the optimal optimization path. Step (2) Initialize the population of block chain trusted big data detection algorithm. The initial route of ants from the distribution center of feature distribution is set up, and global update is carried out according to pheromones. Step (3) Calculate the objective function and the optimal solution of this cycle, and combine ant colony optimization to avoid the algorithm falling into local optimization. Step (4) Using CRO to find out the best solution for high-quality extraction of education resources.

4 Simulation Experiment Analysis In order to verify the application performance of the method in realizing high-quality extraction of education resources, simulation experiments are carried out and Matlab is used for simulation analysis. It is assumed that ant colony performs optimal scheduling path search range: −500 ≤ x d ≤ 500. The optimal value for global search of education resources is min(f6 ) = f6 (0, 0, . . . , 0) = 0.83, and other relevant parameters are set as follows: q0 = 0.6, α = 1, β = 2, λ = 1, γ = 1, σ = 1, ρ = 0.2, ξ = 0.1, m = 50, NCmax = 100, Q = 5. The three optimized paths for high-quality extraction of available education resources are as follows: 1→4→8→4→0, 0→12→3→5→0, 0→4→6→0. Maximum load distribution of education resources: 200, 500, 58. See Table 1 for safety evaluation values between education resources emergency center and dispatching point.

High-Quality Extraction Method of Education Resources

93

Table 1. Safety evaluation values between education resource emergency center and dispatching point sij 0 1 2 3 4 5 6 7 8 0

– 6 7 4 5 6 7 5 5

1

4 – 6 5 6 6 6 4 8

2

6 8 – 4 7 5 4 4 4

3

8 4 5 – 6 4 3 3 6

4

4 3 3 7 – 7 4 4 7

5

7 4 4 3 5 – 5 5 3

6

9 7 5 4 4 5 – 6 4

7

5 4 7 7 6 4 7 – 6

8

4 3 5 4 3 5 6 5 –

According to the above parameter settings, a statistical analysis model for highquality extraction of education resources is established, and the high-quality extraction of education resources and adaptive optimization control are carried out by using autocorrelation quantitative feature analysis and fuzzy information clustering methods, thus obtaining the final optimization path distribution as shown in Fig. 1. 1 0.8 0.6

Amplitude/V

0.4 0.2 0 -0.2

-0.4 -0.6 -0.8 -1

0

200

400

600 Time/s

800

1000

1200

Fig. 1. Optimal path for high quality extraction of education resources

According to the optimization path of Fig. 1, the convergence judgment of education resource scheduling is carried out, and the obtained result is shown in Fig. 2. It can be seen from Fig. 2 that the algorithm in this paper achieves the optimal convergence at about 200, which is far faster than other algorithms, and the wireless convergence is close to 200, which shows that the method has good convergence for educational resource scheduling, ensures the fast and accurate analysis data, and can prevent the emergence of local optimal solutions.

94

H. Zhang et al. 3

FCM PSO QPSO Proposed method

Convergence value

2.5 2 1.5 1 0.5 0

0

200

400 Iterations

600

800

Fig. 2. Convergence curve of education resource scheduling

The higher the number of iterations, the less the corresponding time. The criteria of response time are as follows: (1) Within 3 s, the page responds to the user and displays something, which can be considered as “very good”; (2) Within 3–5 s, the page responds to the user and displays something, which can be considered as “good”; (3) Within 5–10 s, the page responds to the user and displays something, which can be regarded as “reluctantly accepted”; (4) More than 10 s makes people a little impatient, and users are likely not to continue to wait. See Table 2 for comparison results. Table 2. Comparison of algorithm results Algorithm

Parametric performance

Experiment 1

Experiment 2

Experiment 3

PSO

Number of iterations

32

18

16

Response time (Unite: seconds)

1.12

13.54

5.56

Number of iterations

14

18

28

Response time (Unite: seconds)

1.36

13.54

15.53

Number of iterations

3

9

11

Response time (Unite: seconds)

1.03

1.26

8.12

Colony

Proposed method

Table 2 analysis shows that compared with other methods, the maximum response time of this method is 1.36 s, which is far lower than other methods, indicating that the optimization time of extracting high-quality education resources by this method

High-Quality Extraction Method of Education Resources

95

is relatively short, indicating that the response time is well guaranteed, the extraction efficiency of government resources is improved, and the emergency ability of platform search is ensured.

5 Conclusions Combining information fusion and big data analysis technology, the optimization design of high-quality extraction scheme of education resources is carried out. This paper proposes a high-quality extraction scheme of education resources based on block chain trusted big data. Linear programming scheme is adopted to obtain the constraints of the shortest extraction distance, the lowest cost and the largest safety factor of highquality extraction of education resources. Autocorrelation quantitative feature analysis and fuzzy information clustering method are adopted to carry out high-quality extraction and adaptive optimization control of education resources. According to the relevance of block chain trusted big data distribution, the game control of high-quality extraction of education resources is carried out. Combining with the concentration of block chain trusted big data distribution information, the fitness level and equilibrium level of high-quality extraction of education resources are analyzed. The root game equilibrium optimization algorithm realizes the optimization of high-quality extraction of education resources. The analysis shows that the optimization ability of high-quality extraction of education resources by the method in this paper is better, the convergence of the scheduling process is stronger, the optimization scheduling and access ability of education resources are improved, and the consumption time is shorter.

References 1. Huang, S.C., Liu, Y.: Classification algorithm for noisy and dynamic data stream. J. Jiangsu Univ. Sci. Technol. (Nat. Sci. Edn.) 30(3), 281–285 (2016) 2. Sun, B., Wang, J.D., Chen, H.Y., et al.: Diversity measures inensemble learning. Control Dec. 29(3), 385–395 (2014) 3. Du, N.B., Zhan, J.F., Zhao, M., et al.: Spatio-temporal data index model of moving objects on fixed networks using HBase. In: Proceedings of the 2015 IEEE International Conference on Computational Intelligence & Communication Technology. IEEE, Piscataway, NJ, vol. 3, issue 3, pp. 247–251 (2015) 4. Sahu, P.K., Manna, K., Shah, N., et al.: Extending Kernighan-Lin partitioning heuristic for application mapping onto network-on-chip. J. Syst. Architect. 60(7), 562–578 (2014) 5. Komai, Y., Nguyen, D.H., Hara, T., et al.: KNN search utilizing index of the minimum road travel time in time-dependent road networks. In: Proceedings of the 2014 IEEE 33rd International Symposium on Reliable Distributed Systems Workshops. IEEE, Piscataway, NJ, vol. 32, issue 7, pp. 131–137(2014) 6. Ke, S.N., Gong, J., Li, S.N., et al.: A hybrid spatio-temporal data indexing method for trajectory databases. Sensors 14(7), 12990–13005 (2014) 7. Lin, J.M., Ban, W.J., Wang, J.Y., et al.: Query optimization for distributed database based on parallel genetic algorithm and max-min ant system. J. Comput. Appl. 36(3), 675–680 (2016) 8. Zhou, X.P., Zhang, X.F., Zhao, X.N.: Cloud storage performance evaluation research. Comput. Sci. 41(4), 190–194 (2014)

96

H. Zhang et al.

9. Qi, C.Y., Li, Z.H., Zhang, X., et al.: The research of cloud storage system performance evaluation. J. Comput. Res. Dev. 51(1), 223–228 (2014) 10. Mohan, B., Govardhan, A.: Online aggregation using MapReduce in MongoDB. Int. J. Adv. Res. Comput. Sci. Softw. Eng. 3(9), 1157–1165 (2013) 11. Guo, H.P., Zhou, J., Wu, C.A., Fan, M.: k-nearest neighbor classification method for classimbalanced problem. J. Comput. Appl. 38(4), 955–959 (2018) 12. Lu, C., Sheng, W., Han, Y., et al.: Phase-only pattern synthesis based on gradient-descent optimization. J. Syst. Eng. Electron. 27(2), 297–307 (2016) 13. Sun, Y.Z., Fan, W.R., Zhang, S.Q., et al.: Resource allocation algorithm for dense D2D network based on graph coloring. Comput. Eng. 45(2), 26–31 (2019) 14. Zhou, W., Luo, J.J., Jin, K., et al.: Particle swarm and differential evolution fusion algorithm based on fuzzy Gauss learning strategy. J. Comput. Appl. 37(9), 2536–2540 (2017) 15. Meng, X.W., Liu, S.D., Zhang, Y.J., et al.: Research on social recommender systems. J. Softw. 26(6), 1356–1372 (2015)

Application of Open-Source Software in Knowledge Graph Construction Qianqian Cao1,2 and Bo Zhao1(B) 1 Yunnan Normal University, Kunming 650500, YN, China

[email protected], [email protected] 2 Key Laboratory of Educational Information for Nationalities, Ministry of Education,

Kunming 650500, YN, China

Abstract. Knowledge graph (KG), as a new type of knowledge representation, has gained much attention in knowledge engineering. It is difficult for researchers to construct a high-quality KG. Open-source software (OSS) has been slightly used for the knowledge graph construction, which provide an easier way for researchers to development KG quickly. In this work, we discuss briefly the process of KGC and involved techniques at first. This review also summarizes several OSSs available on the web, and their main functions and features, etc. We hope this work can provide some useful reference for knowledge graph construction. Keywords: Knowledge graph · Open-source software · Knowledge graph construction

1 Introduction Knowledge graph (KG) is helpful to organize massive information efficiently. Although KG is applied in semantic search [1], intelligent question-answering [2] and decisionmaking widely, it’s too difficult for many people to construct knowledge graph and its method of automatic construction is not mature yet. However, there are a great many outstanding open-source softwares available and each has its own unique features. It is a key for developers how to choose an appropriate open-source software (OSS) to construct the knowledge graph. In the paper, we have analyzed and compared the characteristics of open-source tools which are applied in the process of knowledge graph construction (KGC), in order to give researchers a reference. Firstly, we discuss briefly the general process of KG from the relevant literature. Secondly, some of the current OSSs for knowledge graph construction are analyzed in Sect. 3. Finally, the conclusion is given.

2 The Process of Knowledge Graph Construction The term “knowledge graph” was first proposed to improve the quality of information retrieval by Google [3] in 2012. Since then, the term has been widely used without a © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 97–102, 2020. https://doi.org/10.1007/978-3-030-63955-6_9

98

Q. Cao and B. Zhao

specific definition and representation of the knowledge based on graph could be considered a knowledge graph. In this paper, KG is a semantic graph consisting of vertices (or nodes) and edges, while the vertices represent concepts or entities and the edges represent relationships between the entities [4]. It able to organize information in an easy-to-maintain, easy-to-understand and easy-to-use method. It has become prevalent in education during these years. The process of knowledge graph construction includes 3 aspects: knowledge acquisition, knowledge fusion, knowledge processing, which will be discussed as follows. 2.1 Knowledge Acquisition (KA) KA is used to mine entities, entity attributes and relationships between the entities from structured, semi-structured, even unstructured data sources. There are mainly three corresponding processing methods for three types of data respectively when knowledge is acquired. For structured data, D2R [5] (an XML-based language) is used to map it to RDF [6] schema [7] (a type of hierarchy which is used to define types and possible relations). RDF (Resource Description Framework) is generally used to formalizes structured information and present it graphically. Semi-structured data (e.g. encyclopedic data, web data) can be gained automatically by using an unsupervised clustering algorithm. For unstructured data, the information extraction technique is mainly used to extract entities, entity attributes and relationships between the entities. Open information extraction (OIE) [8] is the promising field in information extraction. OIE is based on linguistic model and machine learning algorithm, which is used to extract information from open domain. However, the OIE technique might lead to the lower precision and lower recall rate. Through the process of KA, entities, attributes and relations between entities are obtained from structured, semi-structured and unstructured data sources. 2.2 Knowledge Fusion (KF) Through knowledge acquisition, it might contains lots of redundant and erroneous information. KF focuses on re-cleaning and integrating those extracted results, such as improving the connection density (the distance in the KG) of entities/relations and making entity relations to present definite logic and levels. It mainly includes such task as entity linking and knowledge merger. Entity linking focuses on the process of entity disambiguation and entity resolution. Entity disambiguation is used to disambiguate a polysemous entity mention or infer that two different mentions are the same entity, and entity resolution is used to identify and link different manifestations of the same real world entities in various records. Then, entities are mapped from an input text to corresponding entities in a target knowledge base. So, entity linking is used to link the entities extracted from semi-structured and unstructured data sources to knowledge graph. Knowledge merger is used to link the entities extracted from structured data sources (e.g. external knowledge base data or relative database data) to knowledge graph. Through knowledge fusion, the ambiguity of the concepts is eliminated and redundant and wrong concepts are eliminated.

Application of OSS in Knowledge Graph Construction

99

2.3 Knowledge Processing (KP) The result provided by knowledge fusion is not yet a knowledge graph. Knowledge processing is used to transforming the result into the more structured and networked knowledge architecture. KP mainly includes two aspects such as schema construction and knowledge reasoning. Schema construction refers to design the classes/concepts, relations, functions, axioms, instances for the knowledge graph. There are three ways of constructing method: artificial, automated and semi-automated construction. Knowledge reasoning aims at generating new knowledge based on the existing knowledge in the knowledge graph through computer reasoning. There are three common ways to reason over knowledge graph: logic-based reasoning, graph-based reasoning, deep learningbased reasoning. Through knowledge reasoning, knowledge graph is further enriched and expanded.

3 On OSS of Knowledge Graph Construction The term “open-source” is derived from the computer software industry. It can literally be interpreted as open-source code. In the paper, we will analyze and discuss the functions and characteristics of open-source software (OSS) in detail as follows. The commonly used OSSs in knowledge graph construction are shown as in Table 1. Table 1. The commonly used OSSs in knowledge graph construction phases. Knowledge acquisition

Knowledge fusion

Knowledge processing

Protégé

/

/

Yes

KAON

/

/

Yes

DeepDive

Yes

/

Yes

GATE

Yes

Yes

/

KnowItAll

Yes

/

/

Dedupe

/

Yes

/

Limes

/

Yes

/

SOFIE

Yes

/

Yes

Protégé [9] is an open-source software developed by the Stanford University School of Bioinformatics Research Center, which is used for schema construction and knowledge reasoning (knowledge processing). It allows pluggable components, plug-ins, which can visualize knowledge and reason. It has friendly graphic interactive interface, uniform style, strong interactivity, adaptability and operatablity. For schema construction, it is the main functions of Protégé, which include the construction of concept classes, relationships, attributes, and instances of the schema. It is designed to hide the underlying details (the schema description language) from your users. It can also perform knowledge-based reasoning by clicking on the “Reasoner” button (on the menu bar)and provide users with reasoning result and explanation in detail.

100

Q. Cao and B. Zhao

KAON [10] is an open-source toolkit developed by the University of Karlsruhe and the Research Center for Information Technologies in Karlsruhe. It is often used to construct schema (knowledge processing). KAON implements the functions of schema construction by using KAON-API. KAON-API provides mechanisms for storing and editing schema, and also has enabled applications to access and process the schema. KAON’s constructing, storing, querying and retrieving schema operations are all performed based on graph. KAON’s graph-based operation is more intuitive and convenient than Protégé in the constructing process. DeepDive [11] is an open-source system developed by Stanford University’s InfoLab laboratory, which is frequently used for information extraction (knowledge acquisition) and knowledge reasoning (knowledge processing). For knowledge acquisition, it uses “weakly supervised learning” algorithm [12] to extract structured relations from unstructured data, and then determines whether a specified relationship exists between entities. For knowledge processing, DeepDive is used to solve knowledge reasoning problems based on factor graphs, which is a type of probabilistic graphical model composed of variables and factors (definition of the relationships between variables in the graph). DeepDive uses probabilistic inference to estimate the probability whether the existing knowledge would be true, then determinate whether it will be retained. As a complete knowledge extraction framework, DeepDive provides users with application code and an inference engine. Users only need to write application code and inference rules used during reasoning for their specific tasks [13]. KnowItAll [14] is another open-source software developed by the Turing Center at the University of Washington, which is used in knowledge acquisition widely. Training was conducted by unsupervised learning [15] and domain-independent operations. It has advanced performance than previous system in the tasks such as pattern learning and subclass extraction by using the search engine to perform extraction from massive Web data. Reverb and OLLIE are the first and second generation information extraction components of KnowItAll respectively. Reverb extracts the binary relationship directly from the English sentence without to pre-specify the relations. OLLIE system can extract information based on the “Syntax Dependency Tree”. It has better extraction results and the more optimized “Long-term Dependency” effect than Reverb based on text sequence The distinctive feature of KnowItAll is its use of bootstrapping method that does not require any manually tagged training sentences [16]. GATE [17] (General Architecture for Text Engineering) is an open-source knowledge extraction system developed by the University of Sheffield, UK, which is capable of solving almost all problems encountered when processing text. Therefore, it can be used in information extraction (knowledge acquisition) and entity disambiguation (knowledge fusion). For knowledge acquisition, GATE offers users the functions such as data processing, rule definition and entity disambiguation through its natural language processing components, which lays the foundation for information extraction. Each component has the open interface, so that they can be called by other systems freely. Limes [18] is an open-source entity linking framework developed by the DICE (Data Science) group at Paderborn University. More specifically, it is a link discovery framework for metric spaces (a set together with a metric on the set.), which is used in

Application of OSS in Knowledge Graph Construction

101

knowledge fusion. It uses statistics, prefix suffixes, and location filtering to calculate the similarity rate between the entities, while the “entity pairs” mismatched will be filtered out. The specific characteristic of Limes is that it allows users to configure the rules of entity resolution flexibly. Dedupe [19] is an open-source python library developed by Gregg Forest and Derek Ederalso, which is used for knowledge fusion. It uses machine learning to perform fuzzy matching, deduplication and entity resolution quickly on structured data. Specifically, users only need to label a small amount of data selected during the calculation process. All labeled data is clustered and grouped by a compound clustering method. Then the duplicate labeled data will be removed based on the calculation of similarity features [20] and machine learning models [21]. A salient feature of Dedupe is that it supports users using user-defined data types to label the data which are used for training models. SOFIE is an open-source software about knowledge processing developed by the Max Planck Institute, which can also be used to extract information (knowledge acquisition). For entity fusion, it can be used to parse natural language documents, extract knowledge from them and link the knowledge into an existing schema, and perform disambiguation based on logical reasoning. For knowledge acquisition, it processes the document, splits the document into short strings. SOFIE’s [22] main algorithm is completely source-independent. SOFIE’s performance could be further boosted by customizing its rules to specific types of input corpora. With appropriate rules, SOFIE could potentially accommodate other IE paradigms within its unified framework.

4 Conclusion In this article, we have reviewed how the knowledge graph is constructed. We have discussed several the main OSSs available and the usage of these tools in the knowledge graph construction. As previously discussed, there are lots of open-source tools for constructing knowledge graph. We argue that different tool has its own features, such as Dedupe is more suitable to process structured data, DeepDive has better performance when it is used to extract sophisticated relationships between entities, SOFIE integrates information extraction with logical reasoning firstly. To sum up, KGs have arisen as one of the most important tool for knowledge representation, storage and organization. It is hope that more open-source softwares will be developed to provide researchers more convenient ways to build, update, retrieve, and maintain knowledge graph. Acknowledgments. The research is supported by a National Nature Science Fund Project (No. 61967015), and specific project of teacher education of Yunnan Province education science planning (Union of higher education teachers) (GJZ171802).

References 1. Grainger, T., AlJadda, K., Korayem, M., et al.: The Semantic Knowledge Graph: A compact, auto-generated model for real-time traversal and ranking of any relationship within a domain. In: 2016 IEEE International Conference on Data Science and Advanced Analytics (DSAA), pp. 420–429. IEEE (2016)

102

Q. Cao and B. Zhao

2. Zhang, Y., Dai, H., Kozareva, Z., et al.: Variational reasoning for question answering with knowledge graph. In: Thirty-Second AAAI Conference on Artificial Intelligence, AAAI (2018) 3. Amit, S.: Introducing the Knowledge Graph: things, not strings http://googleblog.blogspot. co.uk/2012/05/introducing-knowledge-graph-things-not.html. Accessed 1 March 2020 4. Yan, J., Wang, C., Cheng, W., Gao, M., Zhou, A.: A retrospective of knowledge graphs. Front. Comput. Sci. 12(1), 55–74 (2018). https://doi.org/10.1007/s11704-016-5228-9 5. Bizer, C.: D2r map a database to rdf mapping language (2003) 6. RDF Working Group RDF. http://www.w3.org/RDF. Accessed 1 March 2020 7. Paulheim, H.: Knowledge graph refinement: a survey of approaches and evaluation methods. Semant. Web 8(3), 489–508 (2017) 8. Banko, M., Cafarella, M. J., Soderland, S., et al.: Open information extraction for the web. In: Proceedings of the 20th International Joint Conference on Artificial Intelligence, pp. 2670– 2676. New York (2007) 9. Musen, M.A.: The protégé project: a look back and a look forward. AI matters 1(4), 4–12 (2015) 10. Bozsak, E., Ehrig, M., et al.: KAON — Towards a large scale semantic web. In: Bauknecht, K., Tjoa, A., Quirchmayr, G. (eds.) EC-Web 2002. LNCS, vol. 2455, pp. 304–313. Springer, Heidelberg (2002). https://doi.org/10.1007/3-540-45705-4_32 11. Zhang, C.: Deepdive: A Data Management System For Automatic Knowledge Base Construction. University of Wisconsin-Madison, Madison, Wisconsin (2015) 12. Zhang, Z.: Weakly-supervised relation classification for information extraction. In: Proceedings of the Thirteenth ACM International Conference on Information and Knowledge Management, pp. 581–588 (2004) 13. Mallory, E.K., Zhang, C., Ré, C., Altman, R.B., et al.: Large-scale extraction of gene interactions from full-text literature using DeepDive. Bioinformatics 32(1), 106–113 (2016) 14. Etzioni, O., Cafarella, M., Downey, D., et al.: Unsupervised named-entity extraction from the Web: an experimental study. Artif. Intell. 165(1), 91–134 (2005) 15. Zigoris, P., Eads, D,. Zhang, Y.: Unsupervised learning of tree alignment models for information extraction. In: Sixth IEEE International Conference on Data Mining-Workshops (ICDMW’06), IEEE, pp. 45–49 (2006) 16. Etzioni, O., Cafarella, M., Downey, D., et al.: Web-scale information extraction in knowitall: (preliminary results). In: Proceedings of the 13th International Conference on World Wide Web, pp. 100–110 (2004) 17. The University of Sheffield. GATE Information Extraction. http://gate.ac.uk/ie/. Accessed 2 March 2020 18. Data Science Group at UPB. Limes. https://github.com/dice-group/LIMES. Accessed 2 March 2020 19. Gregg, F., Derek, E. Dedupe, https://github.com/datamade/dedupe, Accessed 3 March 2020 20. Mitra, P., Murthy, C.A., Pal, S.K.: Unsupervised feature selection using feature similarity. IEEE Trans. Pattern Anal. Mach. Intell. 24(3), 301–312 (2002) 21. Wikipedia contributors. Machine learning. https://en.wikipedia.org/w/index.php?title=Mac hine_learning&oldid=944212235. Accessed 6 March 2020 22. Suchanek, F., Sozio, M.,Weikum, G.: Sofie: a self-organizing framework for information extraction. In: Proceedings of the 18th International Conference on World Wide Web, pp. 631– 640 (2009)

Design of Network Security Defense Knowledge Training Management Platform Under Cloud Media Ji-yin Zhou and Chun-rong Zhou(B) Chongqing Vocational College of Transportation, Chongqing 402247, China [email protected], [email protected]

Abstract. With the continuous progress of Internet technology, more and more people are getting information through the network, so the number of cloud media is increasing, which brings great hidden danger to the network security. However, the traditional training management platform is difficult to deal with massive data, which leads to the imbalance of defense knowledge system in the platform and the instability of platform feedback. Therefore, a new network security defense knowledge training management platform is designed. By setting the balance constraint index, optimize the load balance layout of the training management platform; build the service framework of the training management platform, obtain the training user clustering results; set up training courses, realize the design of the network security defense knowledge training management platform. The test results show that compared with the traditional design platform, the stability of the designed platform is higher under the condition of massive network security defense knowledge. It can be seen that the designed platform is more in line with the basic requirements of network security defense. Keywords: Cloud media · Network security defense · Knowledge training management platform

1 Introduction With the development of Internet technology and the rise of social network, the media presents the trend of fragmentation. Everyone has become a media. This media can disseminate information and distribute information. It is displayed in personal blog, micro-blog, space home page, group and so on. At the same time, some enterprises and institutions have similar tendencies in QQ group, website, official micro-blog and WeChat official account, and all kinds of Internet applications, games and websites are showing. Interactive communication function, the collection of human media and these fragmented media, is called cloud media. Network security refers to the use of various technologies and management measures to ensure the normal operation of the network system, so as to ensure the availability, integrity and confidentiality of network data [1].

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 103–114, 2020. https://doi.org/10.1007/978-3-030-63955-6_10

104

J. Zhou and C. Zhou

In response to the above problems, traditional training management platforms have set up key information and developed platform search programs to achieve reliable feedback on network security defense data. However, due to the increasing number of cloud media and the emergence of network attack methods, the amount of data on network security defense knowledge has increased. Traditional platforms have been unable to meet the training requirements. Therefore, a new network security defense knowledge training management platform is proposed. The emergence of this platform makes up for the shortcomings of traditional design platforms, and provides more rigorous technical support for national cyber security education and training [2].

2 Design of Network Security Defense Knowledge Training Management Platform Under Cloud Media 2.1 Set Balance Constraint Indicators According to the amount of network security defense knowledge training tasks, the training management platform needs to balance the distribution and control of these training tasks by setting balance constraint indicators, so as to achieve the optimal layout of training management platform load. The unequal probability task allocation balance is mainly composed of front-end distribution and end service, so a description model is built according to this condition. Assuming that there is no connection between the distributor and the server, that is, the former does not know the load information and task quantity of the latter; at the same time, by default, the end servers can assign tasks to each end server according to a certain probability. Set the number of end servers connected to the management platform to n, and the probability mentioned is represented by the letter m  pn = 1 exists. Let the Poisson flow of the task reach the pn , where 1 ≤ n ≤ N exists, i=1

distributor be γ , then at this time reach the end of the Poisson flow of the server n with the parameter γ pn . At the same time, ensure that the training task time is subject to the general allocation, and match according to the order of queuing rules, and require that the data arrival process and service process are independent of each other [3]. The training task distribution function is known as f (x), and the training task is  divided into N disjoint intervals, that is, [0, ∞) = [u0 , u1 ), [u1 , u2 ), . . . , uN −1 , uN ), where u0 = 0 and uN = ∞. The dispatcher quickly assigns the task uN to the server according to the probability pn . When the server receives the dispatcher’s task, it analyzes the training workload. When un−1 ≤ u < un , the end server places the training task in the queue and waits for service; otherwise, the server returns the task, and the distributor reassigns the task to server n, which satisfies the task assignment condition ui−1 ≤ u < ui . On the basis of setting the assignment conditions, set the control conditions for task assignment. It is known that the capacity of the end server is c, and there is a connection between the front-end distributor and the end server. Let C1 and C2 represent the control programs in the two end servers. When C1 − C2 ≥ λ, the distributor directly assigns tasks to C2 , and the new tasks are in The C2 server waits in line, while C1 does not accept training tasks temporarily; when C2 − C1 ≥ λ, the front-end distributor assigns

Design of Network Security Defense Knowledge Training Management

105

training tasks to C1 , new tasks wait for services in C1 , and C2 temporarily does not accept new assignment tasks; when condition |C2 − C1 | < λ exists At the same time, the distributor distributes system tasks to C1 and C2 at the same time, so that the services of the end server are performed simultaneously. The distribution and control curve of network security defense knowledge training tasks is shown in Fig. 1 below [4]. 1.0 Distribution curve

0.8

Control curve

0.6

Linear index

0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1.0 0

0.2

0.4

0. 6

0.8

1.2

1.0

1.4

1.6

1.8

2.0

index

Fig. 1. Distribution and control of training tasks

According to the above figure, the distribution curve distributes the system task amount evenly, and the control conditions are based on the control law to reasonably control the task allocation amount within a certain range to ensure the distributed parallel computing capability of the distributed management platform. The control conditions are based on the balanced distribution of training tasks, the characteristics of tilted data in the training platform are extracted, and the balance constraint index is set. The setting of this indicator allows the training management platform to run smoothly under normal load conditions, control the degree of data inclination, and make the platform’s processing result of defensive knowledge to be optimal. Suppose the total amount of inclined data in defense knowledge is w, where the data characteristics corresponding to wi are represented by si . The load balancing algorithm is used to calculate and classify the constraint indicators. The indicators in Table 1 below can be used as candidates. The data in the above table includes the obtained values of alternative indicators. An index function is established to calculate the load of each node of the defense knowledge cluster on the distributed platform. Take the following formula as the basic premise of constructing index function: f (wi ) = 1 −

n 

(1 − wi )

(1)

i=1

In the formula: f (wi ) represents the indicator substitution function; combined with a set of non-negative parameter ϕi corresponding to the impact indicator, multiplying

106

J. Zhou and C. Zhou Table 1. Indicators of balance constraints

Serial number

Index name

Get numerical value

1

CPU utilization

87–98%/ms

2

CPU temperature

40–65 °C

3

Memory usage

70–85%

4

Platform steady state response time

25 ms

5

Platform transient response time

12 ms

6

Network rate

55M

7

Disk I/O rate

75–95%

the indicator with the calculation result of the above formula, the obtained indicator evaluation function is: μ(g) = si h(xi )

n 

ϕi fi

(2)

i=1

In the formula: μ(g) indicates the index evaluation function; fi indicates the importance of the corresponding influencing factors to the final target; si indicates the characteristics of the tilted data [5]. According to the above formula, the load constraint indicators of each distributed node in the platform are obtained: σ =

κηg(wi )  −ε ω ln μ(g) +c

(3)

In the formula:σ represents the constraint function on wi ; κ represents the weight coefficient; c represents the supplementary parameter; ε is the balance index; η represents the maximum load of the system; ω represents the load weight of the connection. According to the characteristics of tilt data, the above formula sets a balance constraint index to realize the balance constraint on the load of tilt data connection. Based on the control conditions and constraints set up, a balanced load layout is optimized to realize the balanced control of the inclined data connection load on the training management platform. When the load balance is on the high side, the existing load balance state can be adjusted in time. Assuming that the number of labels of program ti is load is ni , the equilibrium layout probability is obtained according to the balance constraint of formula (4): qi =

ni |n∗ |

(4)

Where: n∗ represents the actual number of loads in the layout area. Then the information entropy D1 can get the maximum value, that is, maxD1 is obtained. According to the results, when the load of each program on the network platform is the same, the index D of the information entropy approaches 1. It should be noted that the index D at

Design of Network Security Defense Knowledge Training Management

107

y10 y9 y8

Ordinate/y

y7 y6 y5 y4 y3 y2 y1 0 0

x1

x2

x3

x4

x6 x5 Abscissa/x

x7

x8

x9

x10

Fig. 2. Optimal load location with less defense knowledge

this time is the label coverage. Obtained under the action of qj . At this time, the amount of knowledge data is small, and the optimal location of load layout is shown in Fig. 2 [6]. According to the above figure, the mathematical model at this time can balance the load when a small amount of data is skewed. It can be seen that under cloud media, the amount of network security defense knowledge is extremely large, so the amount of skewed data also increases. When the amount of skewed data is too large, the load layout needs to readjust the balance. According to the standard particle swarm and cutting function, to a cut for large tilt data processing, data, which originally should tilt left the tilt data to be processed, also bring balance constraint index, combined with particle swarm to optimize the layout, the defense knowledge under the condition of large quantity of the data, the load optimal location are shown in Fig. 3 below [7]. y10 y9 y8

Ordinate/y

y7 y6 y5 y4 y3 y2 y1 0 0

x1

x2

x3

x4

x6 x5 Abscissa/x

x7

x8

x9

x10

Fig. 3. Optimal load location in case of large amount of defense knowledge

108

J. Zhou and C. Zhou

According to the above two sets of optimal position diagrams, it can be known that through the set constraint indicators, constraints are set on the training platform to ensure that the defense knowledge data can maintain a balanced load control balance during the tilt change process and achieve balanced control of the training management platform. 2.2 Build the Service Structure of the Training Management Platform The construction of network security defense knowledge training and management platform under cloud media will focus on online training management and unified data. The four directions of management, diversified training implementation and comprehensive information service are to build a comprehensive training portal platform with better service performance. The specific objectives are as follows: Establish a training management portal, formulate and track training plans and training implementation processes, and achieve unified release of training consulting. Improve the onlineization of the overall training management process, and onlineize most of the offline business. At the same time, an information service monitoring center was established to improve management decision support. Then, it is structured with the management knowledge training thinking of the management platform, providing a unified training service resource platform for network users, integrating teaching video websites and examination system resources, docking with cloud media service centers, and achieving the construction of diversified information targeted services. Portal for integrated information services. Construct a comprehensive training management platform, and form a comprehensive learning management platform service framework with curriculum learning center, teaching management center, information monitoring center, online learning center and quality evaluation center as the main content, as shown in Fig. 4 below [8].

User authentication service

Authentication

Gateway Center

Issuing consultation

Message push service application service

Gateway

Training service

User wechat

Forum services Convenient training

WeChat service

File service

Data retrieval service Decision engine service

Business management Training management center

Training Monitoring Center

Personal information center

Fig. 4. Service framework of training management platform

Design of Network Security Defense Knowledge Training Management

109

Based on the background and goals of platform construction, the scope of platform training task scenarios is divided, as shown in Fig. 4 above. First, the core business scenario is divided into: training services, WeChat services, forum services, and according to three core tasks, training management, training monitoring and personal information center-related business management, while training and forum services provide unified consultation for the portal center Launched, WeChat service provides users with convenient office channels; secondly, the platform authority authentication related tasks are divided into user authentication services; finally, platform training management tasks are divided into: message platform services and file services, and data function tasks are divided into data retrieval services and decision Engine services. Integrate network security defense knowledge training management resources based on the above business divisions. Assume that the set of knowledge requirements for network security defense is X , and randomly select two requirements from them, record them as xa and xb respectively, and find the similarity between them.

τab

n  xai · xbi qi i=1 =   m m   2 2 σ xai xbi i=1

(5)

i=1

Use the above formula to find the degree of similarity between the demand data, and use the best domain value method to achieve fuzzy clustering of user needs: nε  (ε)

xi

=

k=1

(ε)

xki

τab

(6)

(ε)

In the formula: xi represents the clustering result of the i-th sample data under the (ε) influence of ε-type demand; xki represents the dynamic change constant of the demand data under the action of the change factor k [9] According to the service framework of the training management platform, the training service feedback program is set up. According to the above calculation steps, the fuzzy clustering of training user needs is realized. 2.3 Implementation of Training Management Platform Design According to the service framework of the training management platform established above, and combined with the clustering results of training user needs, the design of the network security defense training management platform is realized. Select advanced network security defense personnel to compile training courses, so that the training users can start training and learning according to the process in Table 2 below, and learn the more difficult course content level by level based on the points obtained.

110

J. Zhou and C. Zhou Table 2. Course credit standard Number Course level

Get integral

1

Junior course

100

2

Intermediate course

200

3

Advanced courses

300

4

Extended Curriculum 350

5

Practice operation

400

Users are trained to learn network security defense knowledge in accordance with the sequence of courses in the above table. Only when the course credits are met, can the next stage of course tasks be carried out. The measurement formula of training integral is as follows: I = I1 × I2 + I

(7)

In the formula: I represents the course credit; I1 represents the integral base; I2 represents the integral multiplier; I represents the reward points given during the prescribed course study time, generally between 0.1-0.3 points. The following Table 3 is the standard value table of integral multiplier I2 [10]. Table 3. Integral multiplier standard Grading scale Assessment score

Multiplier

I

More than 90 points 1

I

90–80 points

0.8

III

80–70 points

0.7

IV

70–60 points

0.5

V

Below 60 points

0

Train users to learn network security defense knowledge in accordance with the above-mentioned processes and course management methods. So far, the design of a network security defense knowledge training management platform under cloud media has been realized.

Design of Network Security Defense Knowledge Training Management

111

3 Test Experiments A comparative experiment is proposed to compare the training management platform designed this time with the training management platform under the traditional design to compare the stability of the feedback data of the two platforms. In the experiment, the designed platform was taken as experiment group A, and the platform under traditional design was taken as experiment group B. Specific experimental conclusions were drawn based on experimental test results. Figure 5 below is the test platform built for this experimental test. Computer 1

Computer 2

Route

Computer 3

Fig. 5. Experimental test platform

In the figure, computer 1 and computer 2 are two computers with the same model and configuration, which are respectively used for the platform design of experimental group A and experimental group B. Computer 3 is a test unit, through which the training results of two groups of platforms are obtained. The basic test environment of the above three groups of computers is shown in Table 4 below. Table 4. Computer test environment Computer

Hardware

Software

1

CPU:BTIV2.66G Memory:256G

Windows Server2019

2

CPU:BTIV2.66G Memory:256G

Windows Server2019

3

CPU:SYD2.13G Memory:512G

Windows XP/IE11.0

According to the test environment in the above table, it can be known that the hardware capacity of the three groups of computers is extremely large, which matches

112

J. Zhou and C. Zhou

the huge volume of network security defense knowledge and meets the requirements of this experimental test.

10

Standard value

9.8 Experimental group A

Stability index

9.6

9.4

9.2

9

8.8

8.6 0

10

20

30

40

50

60

70

80

90

100

110

120

Experimental test time/min

(a) Experimental group A test results 10

Standard value

9.8

Stability index

9.6

9.4 Experimental group B 9.2

9

8.8

8.6 0

10

20

30

40

50

60

70

80

90

100

110 120

Experimental test time/min

(b)Test results of experiment group B Fig. 6. Platform test results with 10G network security defense knowledge

The training management platform is designed using two design methods. When the data volume of network security defense knowledge is 10G, the experimental test results are obtained, as shown in Fig. 6 below. According to the curve trend in the figure above, the stability of the platform in the two designs is similar to the standard value. To ensure the universality of the experimental test, the platform under the condition of 50G network security defense knowledge was tested. The comparison results of the experimental test are shown in Fig. 7 below.

Design of Network Security Defense Knowledge Training Management 10

113

Standard value

9.5

Stability index

9 Experimental group A

8.5

8

7.5

7

6.5 0

10

20

30

40

50

60

70

80

90

100

110

120

Experimental test time/min

(a)Test results of group A 10

Standard value

9.5

Stability index

9

8.5

Experimental group B

8

7.5

7

6.5 0

10

20

30

40

50

60

70

80

90

100

110

120

Experimental test time/min

(b)Test results of group B Fig. 7. Platform test results with 50G network security defense knowledge

According to the above figure, under the condition of 50G network security defense knowledge, the designed management platform can still run smoothly; while under the traditional design, the stability of the platform has dropped below 8.5, far below the standard value. Comprehensive experimental test results show that the stability of the design platform is better.

4 Concluding Remarks In oder to improve the network security defense knowledge training management, this paper designs a network security defense knowledge training management platform in the cloud media environment, which solves the problem of poor stability of the traditional design management platform, provides reliable technical support for network security, and has broad application prospects.

114

J. Zhou and C. Zhou

References 1. Shamim Hossain, M., Xu, C.S., Li, Y.T., et al.: Advances in next-generation networking technologies for smart healthcare. IEEE Commun. Mag. 56(4), 14–15 (2018) 2. Lv, Y.Y., Guo, Y.F., Chen, Q., et al.: Active perceptive dynamic scheduling mechanism based on negative feedback. Proc. Comput. Sci. 131(2), 520–524 (2018) 3. Li, Y, F., Le, T., Han, Q., et al. T.: Research Notes: Distributed Shadow for Router Security Defense. Int. J. Softw. Eng. Knowl. Eng. 28(2), 193–206 (2018) 4. Liu, T., Tian, J., Wang, J.Z., et al.: Integrated security threats and defense of cyber-physical systems. Zidonghua Xuebao/Acta Automatica Sinica 45(1), 5–24 (2019) 5. Yu, H.: Platform design of sports meeting management system for regular colleges and universities based on B/S structure. Wirel. Pers. Commun. 102(2), 1223–1232 (2018). https:// doi.org/10.1007/s11277-017-5178-z 6. Emanuela, C., Francesca, C., Pericle, S., et al.: Design and impact of a teacher training course, and attitude change concerning educational robotics. Int. J. Soc. Robot. 10(3), 1–17 (2018) 7. Li, J., Xing, Z., Xing, J., et al.: Design of optimized planning platform of electric boiler with heat storage to enhance wind power consumption. Taiyangneng Xuebao/Acta Energiae Solaris Sinica 39(11), 3270–3276 (2018) 8. Zhi, M.H., Lin, P., Han, H.Y., et al.: Design and implementation of augmented reality cloud platform system for 3D entity objects. Proc. Comput. Sci. 131(5), 108–115 (2018) 9. Tang, Q.Y., Zhou, Y.M., Zeng, P., et al.: Background modeling and coding of surveillance video with variable block size. Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/J. Comput.-Aid. Des. Comput. Graph. 30(1), 138 (2018) 10. He, H., He, T., Peng, G.: Research on network service management platform for long term mechanism of sports in colleges. Wirel. Pers. Commun. 102(2), 1117–1127 (2018). https:// doi.org/10.1007/s11277-017-5146-7

Design of Computer-Aided Course Teaching Control System Based on Supervised Learning Algorithm Chun-rong Zhou(B) and Ji-yin Zhou Chongqing Vocational College of Transportation, Chongqing 402247, China [email protected]

Abstract. With the development and maturity of computer and multimedia technology, computer-aided instruction represents advanced teaching ideas and methods. Based on this, the design of computer-aided course teaching control system based on supervised learning algorithm is proposed. The hardware composition and software design of the system design are introduced. Through the main modules: login and registration module, learning mode selection module, learning content selection module, test module operation. Relying on the constraints of multiple identities in the teaching process by the logical structure of the database, the teaching management platform can control and supervise student learning. In order to better test the effectiveness of the system designed in this paper, a comparative experiment of system control learning and free learning is performed. The results prove that compared with free learning, the learning of computerassisted curriculum teaching control system can better urge students to learn, and the learning effect is better. The teaching control system is superior to the traditional teaching system in improving students ‘interest and motivating students’ learning enthusiasm. Keywords: Supervised learning · Computer aided · Course teaching · Control system

1 Introduction With the rapid development of information technology today, modern information technology has penetrated into all walks of life and has been widely used, and has played an important role in the daily work of various industries [1]. In recent years, with the continuous deepening and extensive development of national education and teaching reform, the role of information technology in teaching reform has become increasingly prominent. The application of information technology has made teaching reform more efficient, the use and sharing of resources more convenient, and has greatly promoted teaching the process of reform. The core of the course teaching assistant system is multimedia computer technology and network technology. It has the diversity of teaching information carriers, the digital © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 115–126, 2020. https://doi.org/10.1007/978-3-030-63955-6_11

116

C. Zhou and J. Zhou

nature of teaching information processing, the hypertext nature of teaching information organization, the interactive nature of teaching information performance, and the diagnosis of the teaching process. Sex and other characteristics. The introduction of it has caused a more profound change in the traditional teaching method. It has a great advantage compared to the traditional teaching method. It can optimize the classroom teaching structure. Use the curriculum teaching auxiliary system. The teaching information load is large and the teaching content is rich, which is conducive to motivating students to learn. Interest, improve the teaching environment, and optimize the structure of classroom teaching [2]. It is conducive to highlighting teaching points and breaking through difficult points to guide and demonstrate important teaching content and experimental phenomena in more ways, so that students are more impressed with this part of the content. Regarding supervised learning algorithms, this paper uses conditional probability to calculate the similarity between two bodies, that is, the similarity in the extracted data structure [3]. First select any body arbitrarily, regard it as a class (this class contains several observations), and record it as A; The observations of other individuals belong to one category, and each category contains only one observation. Calculating the probability between class A and other observations, we get the similarity between class A and other observations.

2 Hardware Composition of Computer-Aided Course Teaching Control System Based on Supervised Learning Algorithm When designing the hardware scheme of this system, according to the feasibility, applicability and scalability of the network technology, the hardware environment uses the now very mature shared 10 BASE-TEthernet Ethernet. Its bandwidth of 10 Mbps can basically meet the needs of high-speed transmission such as video data streams. The shared 10 Mbps Ethernet media has grown from the original thick and thin coaxial cable to the current double-stranded cable. Its technology is quite mature and its performance is very stable. In particular, 10 BASE-T uses a hub to centrally manage the network connections of each computer, which greatly reduces the failure rate, and a single connection will not affect the connection of the remaining network cables. This overcomes the shortcomings that affect the overall situation when using coaxial cables to connect, and the network operation is more reliable. Considering the future expansion situation, the network of this system has some room for expansion. It can be smoothly upgraded to 10 MbSP bandwidth Ethernet, and segment management of the network segment can be realized. This can be used as a second step-by-step implementation. The system server is Digital Equipmenteor-poration’s CelebrisXJ5100 small workgroup server [4]. Configured as dual Pentium loo CPU, 32MECC memory, dual GB-level SCSI interface hard disk, high-speed CPI socket network card, has strong self-care ability and sufficient expansion capacity. The workstations are two EP int ml 20 multimedia workstations and several 486level PCs with different configurations. Both multimedia workstations use PCI network cards, and the remaining models are ISA network cards.

Design of Computer-Aided Course Teaching Control System

117

3 Software Design of Computer-Aided Course Teaching Control System Based on Supervised Learning Algorithm 3.1 Functional Design of Main Modules Login and registration module. The main task of this module is to collect the basic information of the students, but also to let the students estimate their interest and cognitive ability in this course [5]. It should be noted that the interest value, cognitive ability value, and student grade value at this time are preliminary evaluations. The system will continuously adjust this according to the student’s learning situation in the subsequent learning process, and gradually approach the actual situation of the student. To build a more realistic student model. As shown in Figs. 1 and 2.

Fig. 1. Login module

In the registration module, in addition to the interest value and cognitive ability value estimated by students during registration, in addition to being represented by a number between 0 and 100, other methods can be used, such as: interest can be divided into interested, general and uninterested The three levels divide cognitive ability into three levels: strong, normal, and weak. Learning mode selection module. The task of this module is to ➀ evaluate the student’s grade according to their cognitive ability and learning interest [6]. The specific method of evaluation is described above;➁ Let students choose the learning method. There are two ways of learning: one is to learn under the control of the system, and the other is to learn freely without the participation of the system. If the student chooses the system control mode to learn, the system will choose the learning content for the student according to his grade. Test module. Students will be tested after each knowledge point is completed. The forms of the test questions are multiple-choice questions, blank questions, and yes/no questions.

118

C. Zhou and J. Zhou

Fig. 2. Registration module

3.2 Constraints on the Logical Structure of a Database Based on the analysis of entities and their relationships in the above conceptual structure, data tables and constraints are established in the database, as detailed below. The administrator table is used to store administrator information and administrator authentication without association with other tables [7]. The administrator table is defined as shown in the table (Table 1). Table 1. Administrator’s tables Name

Data type and length

Restrain

Describe

Admin ID

Nvarchar(20)

Major key

Administrator no.

Admin name

Nvarchar(20)

Non-empty, Unique key

Administrator username

Admin Pwd

Nvarchar(20)

Non-empty

Administrator’s password

The teacher table is used to store teacher information and teacher identification, and its data comes from the basic information data of teachers outside the system, and the user name and password, mailbox and so on of the course teaching assistant system are added on the basis of the basic data. All teacher activity record data in the system depend on this table (Table 2).

Design of Computer-Aided Course Teaching Control System

119

Table 2. Teacher tables Name

Data type and length

Restrain

Describe

T-ID

Nvarchar(6)

Major key

Teacher number

T-name

Nvarchar(50)

Non-empty

Name of teacher

T-uname

Nvarchar(20)

Allow empty, Unique key

Teacher registered user name

T-uPwd

Nvarchar(20)

Non-empty

Teacher’s code

T-Sex

Nvarchar(2)

Gender of teachers

T-Dept

Nvarchar(50)

Department of teachers

T-Birth

Date

Teacher’s birthday

T-Email

Nvarchar(50)

Unique key

Teachers’ mail box

The teacher table is shown in the table. The student table is used to store student information and student identity identification. Its data comes from the basic information data of students outside the system [8], and the user name and password, mailbox and so on of the course teaching assistant system are added on the basis of the basic data. Student activity record data in the system are dependent on this table. The student table is shown in the table (Table 3). Table 3. Student tables Name

Data type and length

Restrain

Describe

S-TD

Nvarchar(20)

Major key

Student ID

S-name

Nvarchar(50)

Non-empty

Student names

S-uname

Nvarchar(20)

Allow empty, Unique key

Student registration username

S-uPwd

Nvarchar(20)

Non-empty

Student code

S-Sex

Nvarchar(2)

Student gender

S-Dept

Nvarchar(50)

Student Department

S-Class

Nvarchar(50)

Student classes

S-Birth

Date

Student’s birthday

S-Email

Nvarchar(50)

Unique key

Student mailbox

The resource category table is used to store the classification of course resources in the system. Because the level of the classification can not be determined, the foreign key is set to point to the primary key of this table. The resource category table is shown in Table 4.

120

C. Zhou and J. Zhou Table 4. Resource categories

Name

Data type and length

Restrain

Describe

R-ID

Int

Major key

Resource category number

R-type

Nvarchar(50)

Non-empty

Name of resource category

R-Parent ID

Int

External keys

Category number of resource category

The course resource table is used to store various types of resource document information in each course, such as syllabus, lecture notes, etc. This table only holds the information related to the resource document, such as the file name, path and so on of the specific resource document. The resource category in this table refers to the resource category table; the resource class refers to the lecture task table, and the reference relationship here is implemented by trigger [9]; The resource publisher refers to the teacher table, see the teacher number of the table, the resource release time takes the current system date time as the default value, the number of resource views takes the default value, and the resource title is used to display the resource list. The course resource table is shown in the table (Table 5). Table 5. Course resource table Name

Data type and length

Restrain

Describe

R-ID

Int

Major key

Resource no.

R-path

Nvarchar(50)

Non-empty

Resource file name and storage path

R-type

Int

Non-empty, external keys

Category of resources

R-course

Nvarchar(50)

Non-empty, external keys

Resource-owned courses

R-puber

Nvarchar(6)

Non-empty, external keys

Resource publishing teacher

R-up time

Date time

Windows default

Resource release time

T-Read times

int

Windows default

Number of resource views

T-Explain

Nvarchar(50)

Non-empty

Resource heading

The teaching task table is used to store the teacher’s teaching task information, and as the basis of authority inspection when the teacher operates, so that only the teaching teacher can carry out the operation of course resource management and so on [10]. Among them, the teacher refers to the teacher table, and there is a check constraint in the teaching term to ensure that the value can only be sum. The lecture task table is shown in the table (Table 6).

Design of Computer-Aided Course Teaching Control System

121

Table 6. Teaching assignments Name

Data type and length

Restrain

Describe

M-ID

Int

Major key

Lecture task no.

M-Class

Nvarchar(50)

Non-empty

Classes

M-dept

Nvarchar(50)

Non-empty

Student registration username

M-teacher

Nvarchar(6)

Non-empty, external keys

Instructor

M-course

Nvarchar(50)

Non-empty

Course title

M-Semester

tinyint

Non-empty

School term

M-Year

Nvarchar(20)

Non-empty

School year

The question-and-answer sheet is used to store information about student questions that interact with teachers and students, where the course to which the question belongs refers to the course name in the lecture schedule, a constraint implemented by a trigger; the questioner refers to the student table; and whether the typical question and whether the correct answer are available take default values. The answer sheet is shown in the table (Table 7). Table 7. Questionnaires Name

Data type and length

Restrain

Describe

Q-ID

Int

Major key

Question no.

Q-Title

Nvarchar(100)

Non-empty

Title of question

Q-Course

Nvarchar(50)

Non-empty, external keys

Question-based courses

Q-Puber

Nvarchar(20)

Non-empty, external keys

Question vector

Q-Pub time

Date time

Non-empty

Question time

Q-Content

Nvarchar(MAX)

Non-empty

Specific questions

Q-is Typical

Bit

Typical questions

Q-is Answered

Bit

Is there a correct answer

The Q & A question response form is used to store the information of the students’ responses after the student-teacher interaction, in which the question belongs to the Q & A question table (see the question number in the table); Student ID, and when the respondent is a teacher, the respondent refers to the teacher table. See the teacher’s work ID in the table. This reference is implemented through triggers. The default is whether the responder is a teacher and the answer is correct. The table is shown in Table 8.

122

C. Zhou and J. Zhou Table 8. Replies to questions

Name

Data type and length Restrain

Describe

A-ID

Int

Reply no.

Major key

A-Question ID Int

Non-empty, external keys Reply to question number

A-Content

Nvarchar(MAX)

Non-empty

A-Puber

Nvarchar(20)

Non-empty, external keys Respondents

A-Pub time

Date time

Non-empty

A-is Teacher

Bit

Whether the respondent is a teacher

A-is Correct

Bit

Is the answer correct

Reply Recovery time

The course assignment table is used to store the assignment information assigned by the teacher during the course of the lecture, where the assignment belongs to the course reference task table, see the course name in the table, this constraint is implemented by the trigger; the assignment publisher refers to the teacher table, see the teacher number in the table: the assignment release time defaults to the current system date time. The course assignment sheet is shown in the table (Table 9). Table 9. Course assignment tables Name

Data type and length

Restrain

Describe

H-ID

Int

Major key

Operation no.

H-Detail

Nvarchar(MAX)

Non-empty

Job content

H-Course

Nvarchar(50)

Non-empty, external keys

Courses of assignment

H-Puber

Nvarchar(20)

Non-empty, external keys

promulgator

H-Pub time

Date time

Windows default

Time of release

H-Deadline

Date time

Non-empty

Submission deadline

The student assignment table is used to store the information of the assignment file submitted by the student after completing the assignment assigned by the teacher. The course assignment number to which the submitted assignment refers to the course assignment table, see the assignment number in the table; the submitter refers to the student form, see the table the student ID of the student; the assignment submission time defaults to the current system date and time. The student worksheet is shown in the table (Table 10).

Design of Computer-Aided Course Teaching Control System

123

Table 10. Student homework tables Name

Data type and length

Restrain

Describe

H-ID

Int

Major key

Student assignment no.

H-Parent ID

Int

Non-empty, external keys

Course assignment no.

H-Puber

Nvarchar(20)

Non-empty, external keys

Submitted student number

H-Pub time

Date time

Windows default

Submission time

H-Path

Nvarchar(100)

Non-empty

Submission of job file path and file name

H-Score

Tinyint

Student homework score

3.3 Control and Supervision of Teaching Management Platform The individualized teaching management platform mainly includes: managing the individualized teaching process, the teacher uses this function to realize the customized, modified, updated, deleted, backup and other operations of the individualized teaching process. The definition of the individualized teaching process is the basis of the normal operation of the system and the main data object of the platform; Monitoring students’ learning process, which provides teachers with the function of statistics and analysis of students’ learning data, and presents the results of the analysis to teachers, so that teachers can fully grasp the students’ learning of their own courses taught; Intelligent assisted individualized teaching, mainly from the goal of reducing teachers’ individualized teaching pressure and improving the efficiency of individualized teaching, carries on the intelligent processing to the student’s question and so on, realizes the automatic ranking according to the problem heat; To formulate learning strategies, teachers mainly make rules for the running process of learning flow, and combine with individual teaching process examples to monitor and control the students’ learning process together.

4 Experiment and Analysis 4.1 Experimental Subject Because one of the purposes of evaluation is to compare the advantages and disadvantages of individualized network teaching with traditional teaching, we should set up a control group with traditional network teaching for the experimental group with individualized network teaching, and use how much the effect of individualized network teaching is improved compared with traditional network teaching as the evaluation standard. The method of selecting the subjects is to distribute the questionnaire randomly among the students majoring in computer science in our college. The main content of the survey is the estimated value of the entrance grade and their cognitive ability and their interest in the course of computer programming and operation. from which six students are drawn according to their admission scores and numbered, in which the situation of odd number students and even number is similar. the so-called situation approximation refers to the students’ entrance scores and interest in this course are close, so that the

124

C. Zhou and J. Zhou

odd number group can be used as the experimental group and the even number group as the control group. as shown in Table 11. Table 11. Student enrolment and interest values Number

Enrolment achievement

Students initially estimated cognitive ability

Initial estimated interest of the student

1

511

90

90

2

506

93

95

3

523

100

100

4

487

90

90

5

463

80

80

6

457

75

75

7

423

60

61

Student levels for systematic evaluation B A C D

4.2 Experimental Methods and Results The system provides two ways of learning: system control learning and free learning. When studying under the system control mode, the system will first evaluate the students’ grade according to the students’ initial estimated cognitive ability value and interest value, then select the corresponding learning content for the students, and provide the test questions for the students after each learning point (or section, unit). Therefore, the comprehensive grades are used to adjust students’ learning ability and grade. The use of free learning is exactly the same as the use of traditional online teaching systems (Table 12). Table 12. Experimental results Number Capacity values Final interest values Student levels given Written test results given by the system by the system 1

89.8

2 3

96.6

4 5 6

78.5

90

90

94

95

95

95

100

100

96

85

90

82

85

80

78

75

75

68

Design of Computer-Aided Course Teaching Control System

125

4.3 Analysis of Results The results of both groups were basically normal distribution, among which: the average score of odd group was 69.0, and the pass rate was 70%. The average score of even groups was 60.2 and the pass rate was 60%. Therefore, through learning, both groups basically meet the requirements of the syllabus, but the effect of learning odd groups by systematic control is better than that of even groups by free learning. For the odd group middle school students, the value of learning ability given by the system is basically consistent with the final test results, which shows that the student model established by the system is basically reasonable. For the odd group middle school students, the value of learning ability given by the system is basically consistent with the final test results, which shows that the student model established by the system is basically reasonable.

5 Conclusion A computer aided course teaching control system based on supervised learning algorithm is proposed. Firstly, the system architecture, development environment and the key technologies used are introduced, then the experimental situation of the system is introduced, including the experimental object, experimental method and experimental results, and the experimental results are analyzed. At the same time, the experimental results show that the interest of students in even group has not changed, while the interest of students in odd group has increased. It can be seen that the teaching control system is superior to the traditional teaching system in improving students’ interest and arousing students’ learning enthusiasm. Students make comprehensive evaluation and diagnosis, and provide the basis for the formulation of teaching strategies on the basis of supervised learning algorithms.

References 1. Jia, N., Zhao, C., Jia, W.: Evaluation and analysis of homework in garden computer aided design course. Mod. Agric. Sci. Technol. 23(11), 280–289 (2018) 2. Zheng, J.: Design and realization of English assisted learning system based on computer. Appl. Microcomput. 34(12), 99–101 (2018) 3. Li, M.: Exploration and practice of task - driven case teaching in the teaching reform of translation. High. Educ. J. 12(3), 130–132 (2019) 4. Liu, F., Teng, X., Qu, P.C., et al.: On-line course construction and practice for computer-aided design of control systems based on learning communication. China Educ. Technol. Equip. 45(21), 32–34, 41 (2018) 5. Zhang, H., Wang, X., Wang, H., et al.: Research on the supporting teaching resources construction of transportation courses. J. Hubei Inst. Technol. 35(2), 63–66 (2019) 6. Wang, X.: Design of English assisted teaching system based on personalized recommendation. Microcomput. Appl. 35(5), 35–38 (2019) 7. Zhou, H.: Computer aided teaching system based on personalized recommendation. Technol. Commun. 11(14), 110–111 (2019) 8. Sun, K., Zhang, F., Ma, F.: Design and practice of course teaching case based on project driven simulation. High. Educ. J. 64(21), 101–103 (2019)

126

C. Zhou and J. Zhou

9. Zhang, L., Xie, J., Li, J.: The reform and practice of ideological and political teaching in courses based on internet – taking computer aided design as an example. Time Agric. Mach. 46(6), 86–88 (2019) 10. Wu, L., Yang, P., Wang, X.: Research on software engineering curriculum design reform in competition - driven collaboration. Exp. Sci. Technol. 17(5), 58–63 (2019)

Research on the Finite Element Analysis of the Sealing Property of the Piston Used in Automobile Teaching Hai-liang Liu1 , Da-wei Ding1 , Xin Huan1 , and Fang-yan Yang2(B) 1 Weifang Engineering Vocational College, Qingzhou 262500, China

[email protected] 2 College of Sports and Health, Changsha Medical University, Changsha 410219, China

[email protected]

Abstract. In recent years, the continuous development of the national economy has promoted the popularity of automobiles. In the course of automobile teaching, the tightness of the internal combustion engine and its piston in the teaching vehicle is an important factor determining the safety performance of the vehicle. The traditional method for analyzing the tightness of pistons for teaching often causes corresponding adverse effects due to the poor accuracy of the results. Therefore, this study designs a finite element analysis method for the sealing of pistons for automotive teaching. First analyze the shape and material characteristics of the piston for automotive teaching, and then build a finite element model of the piston for automotive teaching to complete the processing of piston stress load. Based on this, the research on piston sealing performance is realized according to the coupling analysis process. The experimental results show that the accuracy of the method is better than that of the traditional method, and the analysis process is less time-consuming, which fully proves the effectiveness of the method. Keywords: Finite element analysis · Piston · Tightness · Coupling analysis · Finite element model

1 Introduction With the continuous popularization of private cars, more and more people need to receive car driving instruction. In teaching vehicles, the tightness of the internal combustion engine is one of the important factors determining its safety performance, and the piston in the internal combustion engine is one of the decisive factors for the sealing performance of the internal combustion engine [1, 2]. As a key component of internal combustion engine, the sealing property of piston is directly related to the working reliability and durability of high-speed internal combustion engine, and also directly affects the emission performance of internal combustion engine. The top surface of the piston is subject to the transient high temperature of the gas, which makes the top of the piston and even the entire piston very hot, and the temperature field is very unevenly distributed. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 127–139, 2020. https://doi.org/10.1007/978-3-030-63955-6_12

128

H. Liu et al.

The temperature gradients in various parts are different, causing the piston to deform thermally. The surface cracked, causing the gap between the piston and the cylinder liner to be damaged, and even the phenomenon of the piston pulling the cylinder and locking. Therefore, it is necessary to carry out research and analysis on the tightness of the piston, understand the sealing state and comprehensive stress distribution of the piston, and then improve and optimize the piston, improve the sealing ability of the piston, improve its thermal stress distribution, and improve the work of the piston of the special vehicle for automotive teaching. Reliability and improved emissions are important [3]. In the traditional analysis process, the piston is generally regarded as a whole, and the analysis perspective is relatively macro, resulting in the problem of poor accuracy of the analysis results. Therefore, in this study, the finite element method is used to optimize the sealing analysis process. The finite element method is a very effective numerical calculation method, which can calculate the deformation, stress and dynamic characteristics of various structures with irregular geometry, complex load and support. The finite element analysis of the sealing of the piston for automobile teaching can improve the driving safety of the teaching car, and at the same time, it also further protects the personal safety of the coach and the students.

2 Method Design In the traditional process of analyzing the sealing property of the piston used in automobile teaching, the piston will be generally analyzed as a whole, resulting in the low accuracy of the final analysis results. In order to solve this problem, based on the traditional analysis method, this study uses the finite element analysis process to design the piston sealing finite element analysis method for automobile teaching. The design process is as follows (Fig. 1).

Research on the Finite Element Analysis of the Sealing Property

129

Literature research

Characteristic analysis of piston

Analysis of physical characteristics

Material performance analysis

Building finite element model

Calculation of piston internal stress

Analysis of piston leakage

Sealing analysis completed

Fig. 1. Finite element analysis process of piston sealing for automotive teaching

Because the construction and analysis of the finite element model are involved in this design process, in order to ensure the accuracy of the analysis results, it is necessary to pay attention to the size of the segmentation unit when unfolding the segmentation to avoid affecting the accuracy of the analysis results. 2.1 Analysis of the Shape and Material Characteristics of Pistons Used in Automobile Teaching In order to improve the reliability of the analysis results, before constructing the finite element model, the shape and material characteristics of the piston for automobile teaching should be studied. In general, the shape of the internal combustion engine piston of the teaching automobile is a circular piston with a necked mouth, and its material is ZL109 g silicon aluminum alloy [4, 5], At room temperature, the elastic modulus is E = 7000 MPa, Poisson’s ratio is μ = 0.25, density is ρ = 2.5 ∗ 103 kg/m3 , thermal

130

H. Liu et al.

conductivity is λ = 125 W/(m.k), and at 10 °C–200 °C, the linear expansion coefficient of the material is χ = 20.95 ∗ 10−6 /°C, the tensile strength of the material is δa = 265.5 MPa, and the compressive strength is δb = 270.5 MPa. In addition to the above, other performance parameters of the piston shall be considered in the actual analysis process, as shown in Table 1. Table 1. Other parameters of piston for automotive teaching Parameter type

Parameter serial number

Parameter contents

Basic parameters

1

Piston skirt diameter

2

Piston top diameter

3

Piston skirt length

4

Piston height

5

Pin hole diameter

6

Height from center of pin hole to top surface

Internal part parameters

Generally speaking, the working environment in which pistons are located is relatively harsh. Therefore, in order to ensure the safety and stability of internal combustion engines used in automobile teaching, the performance requirements of piston materials are high. When selecting piston materials, the following conditions need to be met: small density, small thermal expansion coefficient, good wear resistance, mechanical properties, thermal conductivity, and good processability. The physical diagram of the piston for automobile teaching is shown in Fig. 2.

Research on the Finite Element Analysis of the Sealing Property

131

Fig. 2. Physical schematic diagram of automobile teaching piston

Through the research of the above piston material characteristics, we can fundamentally improve the design accuracy of the finite element model of automobile piston for teaching. At the same time, the rationality of the finite element analysis process can be improved by studying the material properties. 2.2 Constructing Finite Element Model of Piston for Automobile Teaching The above analysis is used as the basis for constructing the finite element model, and the computer is used to simulate and analyze the piston. In recent years, in addition to the continuous development of computer technology and the widespread use of high-performance computers, some large-scale modeling software and computing software have also been improved in practice and matured. In the finite element analysis software, ANSYS/LS-DYNA has been widely used for its superior economic applicability [6]. ANSYS integrates CAD, CAE and CAM technology, which can meet the user’s requirements in the whole process from design, calculation to manufacturing. For this reason, in the research of finite element analysis of teaching piston, the three-dimensional modeling and finite element mesh generation of piston are carried out by ANSYS software. Considering that the piston is a complex three-dimensional component, the combustion chamber offset at the top, the existence of the piston pin seat and the shape of the internal cavity are extremely complicated, so that the entire piston does not have axisymmetric properties. Therefore, when three-dimensional modeling of the piston, in order to more realistically and objectively reflect the actual working condition of the piston, the root system of the piston was fired as a whole, and the three-dimensional modeling of the piston was strictly performed in accordance with the dimensions of the drawing. Considering comprehensively the influence of calculation accuracy and the

132

H. Liu et al.

influence of calculation scale on the piston, the model needs to be simplified, and the piston solid model diagram is shown in Fig. 3.

Fig. 3. Construction results of finite element model for teaching piston

When the 3D solid model of piston is meshed by finite element method, the tetrahedral element in t-deas software is used to realize the automatic meshing of solid model. For the key parts such as combustion chamber, piston pin seat hole and piston ring groove, the t-deas software is used to select the size and shape of the part at will, so as to better express the surface boundary of the piston [7]. Considering both the number of units and the calculation time, this study uses regular meshing for the main body of the piston, and free meshing for the rest. The number of elements and nodes of the piston mesh can meet the requirements of engineering accuracy without consuming more calculator time. Using the designed model as the basic model of the analysis process, the sealing analysis is completed after the finite element is divided. 2.3 Stress Load Treatment of Piston The piston finite element model obtained by the above design is segmented. Analyzing the working process of the piston, it can be known that when the pressure of the gas of the internal combustion engine of the automobile reaches the maximum, the piston is subjected to the most severe stress and deformation under the condition of stable speed. Therefore, the piston should be selected as the analysis condition at the rated power and the highest burst pressure. Assume the maximum burst pressure Pz = 10.2516 MPa that the piston can withstand. The explosive pressure acts on the top surface of the piston, the surface of the combustion chamber, the shore of the firepower, and the ring groove. The pressure inside

Research on the Finite Element Analysis of the Sealing Property

133

the first ring groove is 0.70 Pz. The pressure between the first ring and the second ring is 0.30 Pz; the pressure inside the second ring is 0.30 Pz. Generally, the inertia force of piston is reciprocating inertia force. The direction of reciprocating inertia force is opposite to that of piston acceleration, and the action line is parallel to the cylinder center line. Omitting the slight deviation between the reciprocating mass center and the cylinder center line, we can think that the action line of the reciprocating inertia force coincides with the cylinder center line. The reciprocating inertia force can be determined by Eq. (1): Ai = −mi v

(1)

In the formula: mi represents the mass of the piston, v represents the acceleration of the piston, and the unit is m/s2 . The comprehensive stress of the piston can be obtained through the calculation of the above formula. In the solution of nonlinear finite element problems, the stress-strain relationship is generally described by the strain energy function. In the analysis of piston performance, the most widely used function is Mooney Rivlin function, and the strain potential energy can be expressed as shown in formula (2): Q = C10 (P1 − 3) + C01 (P2 − 3)

(2)

In the formula: Set the mechanical property constants as C10 and C01 , and set P1 and P2 as the strain invariants of the piston. The mechanical constants C10 and C01 of the piston can be determined by unidirectional tensile and compression tests. However, due to cost and test complexity in practice, the values of C01 and C01 are often determined by the comprehensive material properties of the piston, which can be expressed by formula (3):   C01 lg 6C10 (1 + ) = 0.0158O − 0.524 (3) C10 n the formula: O represents the hardness of the piston material. Therefore, as long as the hardness value of the piston and the ratio of C10 and C01 are determined, the values of C10 and C01 can be determined. In general, when C10 /C01 is 0.30, the analysis result is the most reasonable. In addition, there is an incompressible constant g in the calculation process, and its size represents the compressibility of the material. If the material is completely incompressible, then g = 0. For piston materials, there are: g=

(1 − 2δ) C10 + C01

(4)

In the formula, δ represents the comprehensive hardness of the piston material [8]. Through the above formula, the comprehensive stress inside the piston and the hardness characteristics of the piston are obtained, so as to facilitate the analysis of the sealing performance of the piston.

134

H. Liu et al.

2.4 Coupling Analysis of Piston Sealing Through the research in the above part, the construction of the finite element model of the piston is completed. In this section, the sealing performance of the piston will be analyzed. In order to ensure the validity of the analysis results, the set analysis calculation process is shown in Fig. 4. Reference finite element model

Calculate pressure distribution

Calculation of distributed load

Whether it meets the requirements of piston

N

Y Calculate the range of piston motion

Calculate the leakage of the piston

Fig. 4. Schematic diagram of piston sealing coupling analysis process

Piston tightness analysis is performed according to the analysis process shown in Fig. 4. In this study, the calculation is only performed on the calculation part after the finite element analysis method is referenced. The specific calculation part is as follows: Generally speaking, the reaction force of the piston pin seat acts on the contact surface between the piston pin and the inner circle of the pin hole. According to the prior knowledge, the force is distributed according to the cosine law within the 90° angle above the ring direction, and approximately according to the triangle along the axial direction. The pressure distribution curve is: I 45 W (R − X ) cos(1.5α)dx dα

W =4 0

(5)

0

In the formula: R represents the length of the contact surface between the piston pin and the pin seat hole, X represents the radius of the pin seat circular hole [9, 10], W represents

Research on the Finite Element Analysis of the Sealing Property

135

the combined reaction force acting on the piston pin seat, and the upper surface of the piston pin seat hole is 120° The distributed load at any point a in the angular range is: w = wx (R − x) cos(60 − 1.5α)

(6)

Using the above formula, the distributed load at one point of the piston is obtained. According to the calculation results of distributed load, the movement range of piston is calculated. Set the operating speed of the piston as y, and calculate the range of motion as follows:  2 2μy 2 (7) t = ti = 3 3 ( ddxε )i In the formula, ( ddxε )i represents the drunk gradient force acceptable to the piston. When the movement mode of the piston is uniform, its moving space is half of the range of the above set moving space, then there is the following formula:  1 1 2μy 1 ∗ (8) t = ti = ti = 2 3 3 ( ddxε )i Combining the above formulas, the leakage of the piston is as follows:   2μy0 2μy1 1 ) π dy( d ε − 3 ( dx )i ( ddxε )i

(9)

In the formula, d represents the movement height of the piston. Through the above results, we can know the leakage of the piston, thus reflecting the sealing performance of the piston. So far, the design of the finite element analysis method of piston sealing for automobile teaching has been completed.

3 Simulation Test Experiment and Result Analysis In order to verify the feasibility of the finite element analysis method of piston sealing used in automobile teaching, the experimental test of the method is completed in the form of simulation experiment analysis, and the corresponding experimental results are analyzed and conclusions are drawn. 3.1 Experimental Environment Setup In this experiment, the traditional car engine piston airtightness analysis method and the method in this paper were used to analyze the pre-prepared pistons for car teaching before the experiment, and compare the accuracy of the analysis results of different methods. To ensure the validity of the experiment, the corresponding experimental environment is set up as follows (Fig. 5):

136

H. Liu et al.

Fig. 5. Simulation experiment environment

This experiment was completed on the Visual Studio platform. The experimental environment was an 8-core i7-5960X, a CUP of 4.0 GHz, and the simulation operating system was Matlab R2017. The finite element design software and three-dimensional simulation software are installed in the computer, and the software is used to assist the traditional method and the method of this article to complete the analysis of the experimental sample piston, and to compare the analysis results of different methods. 3.2 Experimental Samples Based on the design of the experimental environment, in order to improve the reliability and validity of the experimental results, the experimental sample piston parameters are set as shown in Table 2. Table 2. Experimental sample parameter setting Parameter number

Content

Set value

1

Piston material

Eutectic silicon aluminum alloy

2

Eutectic silicon aluminum alloy

0.5

3

Modulus of elasticity of piston

70 GPa

4

Material density

3000 kg/m3

5

Thermal conductivity

145.50 W/ m*k

6

Coefficient of thermal expansion

21.0*10−6 /°C

7

Piston quality

0.300 kg

8

Piston pin mass

0.150 kg

Research on the Finite Element Analysis of the Sealing Property

137

Set the number of experiments as 100, use the sample parameters shown in Table 2 to design the corresponding test sample piston, and use the method in this paper and the traditional analysis method to analyze it, and compare the differences between the two different methods according to the accuracy of the analysis results. 3.3 Analysis of Results Using the above-mentioned experimental environment, analyze the piston tightness in the automobile teaching vehicle by different methods, and complete the comparison of the analysis accuracy between the traditional method and the method in this paper. The specific experimental results are shown in Table 3. Table 3. Comparison of analysis accuracy of different analysis methods Number of experiments

The traditional method analyzes the accuracy/%

The analytical precision of the method in this paper/%

10

93.56

98.25

20

96.21

96.30

30

94.35

98.45

40

96.26

96.54

50

95.57

96.24

60

95.24

97.56

70

97.21

98.21

80

97.20

98.01

90

92.50

98.98

100

90.35

98.67

By comparing the experimental results shown in Table 3, it can be known that with the continuous increase of the number of experiments, the accuracy of the results of the analysis of the piston tightness of automotive teaching vehicles by different methods is constantly changing, and the analysis accuracy of traditional methods fluctuates greatly. The analysis accuracy of the method in this paper is relatively small. It can be seen from the horizontal comparison that the analysis accuracy of the method in this paper is always higher than the traditional method. From this, it can be shown that the method of this paper has a better analysis performance of piston sealing in automobile teaching vehicles, and the analysis ability is obviously better than the traditional method. On this basis, in order to further verify the application advantages of finite element method of piston sealing in automobile teaching, the analysis process time is taken as the experimental index, and the performance of the method in this paper and the traditional method in the analysis process time is compared, so as to judge the work efficiency of different methods. The experimental results are shown in Fig. 6.

138

H. Liu et al.

According to Fig. 6, although the number of experiments keeps increasing, the analysis process of the method in this paper is always less time-consuming than that of the traditional method. The minimum analysis process only takes 1.3 min, while the analysis process of the traditional method always takes more than 2 min. It can be seen that the finite element method of piston sealing designed in this paper is more efficient.

Number of experiments/min

3 2.5 2 1.5 1

The traditional method The method in this paper

0.5

20

80 40 60 Number of experiments/ one

100

Fig. 6. Comparison of time consumption of different analysis methods

4 Concluding Remarks In this study, ANSYS software is used to analyze the sealing performance of the piston used in automobile teaching, and the following conclusions are drawn: When using the 3D drawing software Pro/E to establish the geometric model required for finite element analysis, the requirements of the finite element analysis method for the geometric model must be carefully considered. Some subtle structures that do not affect the analysis results should be ignored. Difficulties in meshing and calculation. It is very important for the safety of teaching vehicles that the sealing of piston used in automobile teaching. Therefore, in the future production design, the sealing of piston should be effectively increased.

References 1. Haojie, H.: Finite element analysis of connecting rod of reciprocating piston compressor. Chem. Eng. Des. Commun. 44(11), 97–103 (2018) 2. Yongchun, Y., Ting, Y., Dailong, S., Shuzhan, B., Guihua, W.: Finite element analysis of temperature field and stress field of diesel engine piston under different working conditions. Intern. Combust. Engine PowerPlant 35(05), 26–32+64 (2018) 3. Jihong, S.: Failure analysis of engine piston seal ring. Technol. Innov. Appl. (36), 105–107 (2019)

Research on the Finite Element Analysis of the Sealing Property

139

4. Junqing, Z., Mengmeng, L., Xudong, Z., Xvdong, Z., Liguo, M.: Thermal analysis of a diesel forged steel piston based on FEA. Intern. Combust. Engine PowerPlant 35(02), 77–81 (2018) 5. Xu, L., Guoyu, C., Yuan, L., Bingdong, G.: Thermal stress analysis of piston based on APDL language. J. Tianjin Univ. Technol. 35(04), 14–18 (2019) 6. Junhong, Z., et al.: Research on the piston’s transient stress considering intake effect and non-linear contact. J. Xi’an Jiaotong Univ. 53(05), 16–23+122 (2019) 7. Pukai, W., Qi, K., Lijun, H., Panpan, H., Yi, D.: Calculation and analysis of friction heat generation for piston group based on finite element method. Veh. Engine 03, 53–58 (2018) 8. Yongtao, H.: Cause analysis and solution of vibration of piston compressor. Compressor Technol. (06), 50–53 (2019) 9. Haoquan, W.: Analysis of surface protection and wear resistance of piston in internal combustion engine of automobile. Plant Maintenance Eng. (04), 163–165 (2019) 10. Chunyang, X., Xinchun, C., Mengxue, B.: Force analysis of piston in pumping system based on ABAQUS. Constr. Mach. Equip. 50(01), 22–26+6–7 (2019)

Research on Multi-agent Robot Behavior Learning Based on Fuzzy Neural Network Jun-ru Wang(B) Chongqing Vocational College of Transportation, Chongqing 402247, China [email protected]

Abstract. Human behavior or motion is diverse and complex. In order to design a robot with better sensitivity and better performance, it is necessary to set up multiple control nodes to facilitate the control robot to imitate the trajectory of human motion. However, due to the traditional robot behavior learning method, the control nodes are relatively single, and there is no clear behavior target node for the reference object, which leads to a large deviation in the robot behavior learning trajectory. Therefore, a fuzzy neural network-based Multi-agent robot behavior learning method. This method is based on the fuzzy neural network to control the behavior of robot. By optimizing the learning parameters of robot behavior, it can enhance the search program of behavior learning. According to the multi-level robot behavior learning model, it can identify the master-slave target of reference object and realize a more accurate multi-agent robot behavior learning method. The test results show that compared with the traditional method, the behavior trajectory of the proposed method is basically consistent with the behavior trajectory of the reference object. It can be seen that the method has better performance and meets the research requirements at the current stage. Keywords: Keywords fuzzy neural network · Multi-agent robot · Behavior learning

1 Preface With the continuous improvement of science and technology, multi-agent robots appear. Traditional behavior learning methods, through the robot’s intelligent central control unit, simulate human movement and activity. However, with many tests, it is found that the behavior learning of this method has a large deviation, and the robot’s behavior trajectory does not conform to the actual situation. Therefore, a multi-agent robot behavior learning method based on fuzzy neural network is studied to solve this problem. This method can achieve more accurate behavior learning by strengthening the setting of control node and defining the master-slave motion node of reference target. This method not only enhances the working efficiency of multi-agent robot, but also provides a research idea for the development of science and technology [1, 2].

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 140–151, 2020. https://doi.org/10.1007/978-3-030-63955-6_13

Research on Multi-agent Robot Behavior Learning

141

2 Multi-agent Robot Behavior Learning Method Based on Fuzzy Neural Network 2.1 Robot Behavior Control Based on Fuzzy Neural Network According to the fuzzy theory of fuzzy neural network, the behavior of robot under the control of fuzzy neural network model is studied, The robot’s fuzzy rule i = (1, 2, . . . , n) is of the form: If β1 (t) is Ki1 and βr (t) is reliable data in the Kir set, then: yˆ (t) = − βi (t)Di y(t) + (Wi + Wi )f (y(t)) + (Ui + Ui ) f (y(t − τ )) + I + Vi u(t)

(1)

Where: βi (t) represents the fuzzy set, i = 1, 2, . . . , r represents the known variable, and r is the number of fuzzy rules. u(t) ∈ M p represents the input value of the input robot; matrix Di = diag{di1 .di2 , . . . , din } is a positive definite diagonal matrix, and Wi , Ui ∈ M n×n and Vi ∈ M n×p are known constant matrices. The matrices Wi and Ui are parameter uncertainties and satisfy the following conditions:   (2) [Wi , Ui ] = Ki F(t) Q1i , Q2i In the formula: Q1i , Q2i , Ki represent known constant matrices of appropriate dimensions. Gaussian fuzzy neural network is a fuzzy neural network composed of product reasoning rules, single-valued fuzzy generators, Gaussian membership functions, and central average anti-fuzzifier [3]. From this, the output of a Gaussian fuzzy neural network can be given as: 

y(t) =

m 

ωi [β(t)] − Di y(t) + (Wi + Wi )f (y(t))

i=1

+ (Ui + Ui )f (y(t − τ )) + I + Vi u(t)/

m 

ωi [β(t)]

(3)

i=1

According to the formula (1-3), the output results are as follows: y(t) =

n 

σi [β(t)] · λi /

i=1

n 

[Wi , Ui ]

(4)

i=1

formula:λi represents the membership of state β(t) to fuzzy set Kir in rule i, and the following conditions are met: λi ≥ 0,

m 

σi [β(t)] = 1

(5)

i=1

In the formula: σi represents the network control parameter of rule i. The control program is set up according to the above formula to implement a fuzzy neural network model to control the behavior learning mode of a multi-agent robot [4].

142

J. Wang

2.2 Optimization of Robot Behavior Learning Parameters The fuzzy neural network based on behavior learning is a simple and effective network structure for multi-agent robot behavior learning, but the fixed behavior control parameters are difficult to adapt to the changing external environment. Therefore, particle swarm optimization algorithm is used to modify the control nodes in the network through the implicit parallel operation ability and good adaptive ability. At the same time, the control nodes in the network also provide a better population for the particle swarm optimization algorithm, so that the algorithm can carry out more biased control node search. In the problem of using particle swarm optimization algorithm to optimize the control parameters of robot behavior learning, each particle is regarded as an independent individual in the search space, and the position vector of each particle corresponds to the solution of an optimization problem. Each particle has a speed that determines the direction and speed of the robot’s behavior. The performance of all particles is evaluated by the fitness value of an optimized function [5]. Position of the i particle can be expressed as a vector wi = (wi1 , wi2 , . . . , wiD ), and the speed of the behavior can be expressed as a vector si = (si1 , si2 , . . . , siD ). In the process of controlling the behavior of the robot, the optimal position found by the i th particle is pi = (pi1 , pi2 , . . . , piD ), and the optimal position of all the control nodes obtained by the particle is pk = (pk1 , pk2 , . . . , pkD ). Each particle updates its speed and position according to the following formula:   si (n + 1) = si (n) + x1 b1 pi − wi (n)   + x2 b2 pk − wi (n) wi (n + 1) = wi (n) + si (n)

(6)

In the formula: i = 1, 2, . . . , m represents different particles; n represents the number of iterations, that is, the number of positioning steps of the particles; x1 and x2 represent acceleration coefficients or learning factors, respectively; b1 and b2 represent random numbers with a variation range of [0, 1]. According to the update result of the above formula, set the optimization process of the robot behavior learning control parameters, as shown in Fig. 1 below. Given the appropriate optimization weight, it can balance the local and global search ability, so as to reduce the number of iterations, so as to control the robot behavior learning faster. Generally, the smaller one can improve the local search ability; the larger one can speed up the convergence speed, and the dynamic one can obtain better optimization results than a fixed value. Therefore, adjusting the size can achieve the goal of balancing the local search control ability and convergence speed [6]. There are two ways to achieve dynamic changes: one is to change linearly in the search process of the algorithm; the other is to dynamically change according to a measure function corresponding to the algorithm performance. This optimization uses the first method, and its expression is as follows: ωmax − ωmin (7) ωˆ i = ωmax − n nmax · si (n + 1) In the formula: ωmax and ωmin represent the maximum and minimum inertia weights; n represents the current number of iteration steps; nmax represents the total number

Research on Multi-agent Robot Behavior Learning

143

Fig. 1. Parameter optimization process of particle swarm optimization algorithm

of iterations. The above formula is used to optimize the control parameters of robot behavior learning and enhance the standard and reliability of behavior learning. 2.3 Establishing a Multi-level Robot Behavior Learning Model According to the optimized parameters, a multi-level robot behavior learning model is established to control the multi-agent robot behavior learning trajectory. The hierarchy of learning model refers to the fact that the actual learning program can be divided into several subprograms, which can be subdivided into several lower level subprograms. Subprograms of the same level often have similar functions. Complex programs are characterized by their emergence, which reflects the non-additive and holistic nature of the program from low-level features to high-level features, which are unique to high-level but not at low-level. Different levels of programs reflect the emergence of different properties. High level subprograms have more complex forms of motion and program characteristics. Complex programs are hierarchical, which is an important theoretical basis for establishing behavioral models. Analytic Hierarchy Process provides a mathematical method for quantitative analysis for program analysis and design. When modeling a complex program, the program is first decomposed into several layers, and then the organization is integrated from the lower layer to the higher layer to form the whole program, and a program hierarchy model is established [7].

144

J. Wang

For general decision-making problems, it can usually be described by a three-level structure. The highest level is the overall goal of the program, and the coarse-grained quantization unit is used to describe the system characteristics. The quantization unit describes the model of each subroutine; the lowest layer is the various solutions used to solve the problem, and the fine-grained quantization unit is used to describe the lowestlevel subroutine model. The layers are organized through association variables to form a hierarchical model of complex programs, as shown in Fig. 2.

Fig. 2. Hierarchy model

Some of the environments that multi-robot programs face are often unknown, dynamic, and unstructured. At the same time, due to the limitations of the robot’s own sensors and the presence of environmental noise, each robot in the group will encounter each other during the task to some unpredictable triggers. When a robot program performs a learning task, the group behavior of the robot through local interaction will be very complicated, and the evolution process of the robot group behavior is a random process. Therefore, according to the robot group behavior, the multi-level robot behavior learning model is established, mostly using the Markov property and random probability calculation method to derive the group behavior learning model, the fractal modeling method, as shown in Fig. 3 [8]. Because the robot population behavior has Markov characteristics or approximate Markov characteristics, the state of the program at any time depends only on the state of the previous moment, independent of the historical state, and its dynamic characteristics can be represented by the following probability distribution:   (8) Paa = P At+1 = ωˆ i a |ωˆ i at In the formula: Paa is the transition probability from state a to state a ; at is the state at time t; At+1 is the state at time t + 1. Both fractal theory and program hierarchy reveal the multi-level and multidimensional connection between the whole and the part of complex program, while fractal emphasizes the similarity between the whole and the part. Because complex programs have both diversity and self-similarity, fractal theory can be used to model

Research on Multi-agent Robot Behavior Learning

145

Fig. 3. Schematic diagram of robot fractal modeling

complex programs. Fractals provide a new idea and method for modeling complex programs. So far, the establishment of a multi-agent robot behavior learning model has been realized. 2.4 Identify Objectives and Achieve Behavioral Learning Use the behavioral learning model above to identify learning targets, extract target features through local feature point detection and descriptor analysis. This recognition method describes the local features of the relationship between the master and slave feature points, and constructs feature descriptors by using different feature point gradients, color attributes and location relationships. The proposed feature point extraction algorithms have different performances in terms of complexity, repetition, and lighting robustness, and can be divided into strong feature points and weak feature points. Among them, the strong feature point is more stable and robust than the weak feature point, but its algorithm complexity is also higher than the weak feature point. In addition, in the same learning data, the number of strong feature points is usually less than that of weak feature points [9]. The strong feature point P1 (main point) and the weak feature point P2 (auxiliary point) of the learning object are extracted separately. If there are auxiliary point sets in the neighborhood of the main point P1i that exceed the threshold value, this area is a feature detection area. Figure 4 is a schematic diagram of the effect of extracting the feature points from the main points. In the figure, the blue squares are Harris main feature points and the red circles are fast auxiliary points. It can be seen that the number of main points is less than the number of auxiliary points, and there are auxiliary points exceeding the threshold number in this area, so this area is a feature detection domain. The local feature descriptor of the gesture is constructed using the master-slave points. According to the analysis of Fig. 4, there are many attributes of the primary and secondary points in the detection area, including the color information of the primary and secondary points, the gradient, and the relationship between the secondary points and the position of the primary points. Target recognition is carried out by using the

146

J. Wang

Fig. 4. Master-slave point feature detection area (Color figure online)

local descriptor method for discrete sampling, and the selected discrete sampling mode is shown in Fig. 5 below.

Fig. 5. Master-slave point feature attributes

The detection area sampling mode PS = [Pi ] is given, where PS is the local area discrete coordinate point set with the origin (0, 0) as the center. The above image is a sampling mode with a radius of 18 from the center and an interval of 3. The sampling coordinates of each detection area are generated by taking the coordinates of feature points as offsets. In order to achieve rotation invariance   and give the main direction of the x detection area, the sampling mode point Pi = i can be mapped to the corresponding yi main direction position according to the following formula.   cos αP − sin αP (9) Pi = sin αP cos αP

Research on Multi-agent Robot Behavior Learning

147

In the formula: αP represents the main direction of the current region for all strong and weak feature points. The pixel values corresponding to the sampling points in the main direction mode constitute  i  MPS of the region, that is, the number-dimensional feature of the sampling mode point set. To transform the current feature descriptor NPS = qPS  i  into a binary feature KPS = kPS for accelerated matching:

 ⎧ i i ⎪ k q = h , q ¯ k ⎪ PS PS ⎨ PS  (10) 1, (x ≥ y) ⎪ ⎪ ⎩ hk (xi , yi ) = 0, (x < y) In the formula: q¯ PS is the average value of NPS ; xi and yi are the horizontal and vertical coordinates of the mapping node; hk (∗) is the identification function. So far, based on the above, a multi-agent robot behavior learning based on fuzzy neural network has been realized [10].

3 Simulation Test Experiments In order to test the reliability of the proposed robot behavior learning method, a simulation test experiment is proposed to analyze the behavior learning trajectory of the multi-agent robot under the proposed method. In order to make the experimental results have obvious differences, and take the traditional robot behavior learning method as the control target, according to the experimental results, the learning differences between the two methods are analyzed. Based on the simulation of the differential motion of the end of the robotic arm as the basic motion operation, and the combination of linear motion and spiral curve motion as experimental examples, the trajectory generation effect of multi-agent robot behavior learning under the control of the two methods was tested. The combined linear differential motion simulation effect diagram is shown in Fig. 6 below.

Fig. 6. Simulation effect of combined linear differential motion

148

J. Wang

Two methods are used to control the simulation manipulator in the figure above and start to perform the behavior learning task. The Fig. 7 below is the test results of the learning trajectory of the combined linear differential motion under the two learning methods.

Fig. 7. Comparison results of combined linear motion learning trajectories

Research on Multi-agent Robot Behavior Learning

149

According to the above two experimental results, we can see that the behavior learning trajectory of the proposed robot learning method is very similar to that of the reference object; However, there is a large deviation between the behavior learning trajectory of the traditional method and the motion trajectory of the reference object. In order to ensure the universality of the experimental results, the curve motion of the reference object is studied. Figure 8 is a schematic diagram of the curve motion simulation effect.

Fig. 8. Effect diagram of curve motion simulation

Based on the above conditions, Fig. 9 below shows the test results of the learning trajectory of curve movement under the two learning methods. It can be seen from the motion trajectories of the above two groups of pictures that the control effect of the proposed method is better than that of the traditional method. Based on two experimental tests, it can be seen that the behavior learning method of the multi-agent robot studied in this study has better control effect. In order to further verify the effectiveness of this method, the behavior learning trajectory control accuracy of this method and the traditional method is compared, and the comparison results are shown in Fig. 10. According to Fig. 10, the behavior learning trajectory control accuracy of this method is higher than that of traditional behavior learning trajectory control, which shows that the behavior learning trajectory control effect of this method is better.

J. Wang

Fig. 9. Comparison results of curve motion learning trajectories 100

80 Control accuracy /%

150

60

Methods of this paper

40

Traditional method

20 0

5

10

15

20

25

30

Number of experiments / me

Fig. 10. Comparison of control accuracy of behavior learning trajectory

Research on Multi-agent Robot Behavior Learning

151

4 Concluding Remarks The robot behavior learning method proposed in this paper solves the problem of large deviation of learning trajectory of traditional methods by strengthening the control of robot nodes and improving the accuracy of robot operation behavior. However, there is still a certain degree of error in the behavior learning method proposed this time. In future research, the error should be further controlled to a smaller range.

References 1. Jin, J.C., Ma, X.L.: A multi-objective agent-based control approach with application in intelligent traffic signal system. IEEE Trans. Intell. Transp. Syst. 20(10), 3900–3912 (2019) 2. Zhang, F., Li, N., Yuan, R., et al.: Robot path planning algorithm based on reinforcement learning. Huazhong Keji Daxue Xuebao (Ziran Kexue Ban)/J. Huazhong Univ. Sci. Technol. (Nat. Sci. Ed.) 46(12), 65–70 (2018) 3. Baklouti, N., Abraham, A., Alimi, A.M.: A beta basis function interval type-2 fuzzy neural network for time series applications. Eng. Appl. Artif. Intell. 71, 259–274 (2018) 4. Xu, B., Zhang, X.P., Liu, L.Q.: The failure detection method of WSN based on PCA-BDA and fuzzy neural network. Wirel. Pers. Commun. 102(4), 1–11 (2018) 5. Zhang, A., Ma, S., Li, B., et al.: Tracking control of tangential velocity of EEL robot based on iterative learning control. Jiqiren/Robot 40(6), 769–778 (2018) 6. Mizanoor Rahman, S.M.: Cyber-physical-social system between a humanoid robot and a virtual human through a shared platform for adaptive agent ecology. IEEE/CAA J. Automatica Sinica 5(1), 190–203 (2018) 7. Kim, T., Kim, J.-W., Lee, K.: Detection of sleep disordered breathing severity using acoustic biomarker and machine learning techniques. Biomed. Eng. Online 17(1), 16 (2018) 8. Chi, W.Q., Liu, J.D., Hedyeh, R.-T., et al.: Learning-based endovascular navigation through the use of non-rigid registration for collaborative robotic catheterization. Int. J. Comput. Assist. Radiol. Surg. 13(6), 855–864 (2018) 9. Goodfellow, Ian, McDaniel, Patrick, Papernot, Nicolas: Making machine learning robust against adversarial inputs. Commun. ACM 61(7), 56–66 (2018) 10. Krempel, R., Kulkarni, P., Yim, A., et al.: Integrative analysis and machine learning on cancer genomics data using the Cancer Systems Biology Database (CancerSysDB). BMC Bioinform. 19(1), 156 (2018)

Game Model of Regional Education and Economic Development Based on Association Rule Mining Algorithm Shan Wang(B) School of Engineering and Management, Pingxiang University, Pingxiang 337000, China [email protected]

Abstract. Aiming at the problem that the game model studies the relationship between regional education and economic development, it will lead to a certain contradiction, which will lead to a long time to obtain game analysis results. This paper proposes a regional education and economic development game based on association rule mining algorithm model. Collect and process data on regional education and economic development; define the subject of regional education and economic development; use Apriori algorithm to implement association rule mining on the relationship between the two, and realize the game of the relationship between regional education and economic development through the coupling degree function, and thus complete mining based on association rules Algorithmic regional education and economic development game model building. Through comparative experiments, it can be concluded that the proposed game model can more quickly obtain the results of game analysis, and through the analysis of coupled and coordinated development, we can come up with suggestions for coordinated development of higher education and regional economy. Keywords: Association rule mining algorithm · Regional education · Economic development · Coupling degree function · Game model

1 Introduction In the era of the beginning of the knowledge economy, the rapid development of science and technology, and the globalization of economic development, especially China’s implementation of the strategy of rejuvenating the country through science and education and the development of the western region, and China’s accession to the WTO and the establishment of the ASEAN Free Trade Area, this has brought regional economic and social development Opportunities for development have come, but severe challenges have also been raised. The implementation of the strategy of rejuvenating the country through science and education can essentially be attributed to two interconnected aspects: first, economic development and social progress must rely on scientific and technological progress and educational development; second, scientific and technological progress and educational development are always oriented to economic and social progress [1–3]. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 152–164, 2020. https://doi.org/10.1007/978-3-030-63955-6_14

Game Model of Regional Education and Economic Development

153

From the perspective of systems science, if the entire society is regarded as an organic system, then economy and education are two important subsystems. And the coordinated development of education and economy as a complex system, its complexity is mainly reflected in the diversity of its constituent elements, the interrelated dynamics and diversity, and there is a dynamic material cycle within the system and between the system and the environment. Energy flow, information transfer and value addition. A qualitative analysis alone is not enough to grasp the behavioral and functional characteristics of the system as a whole. Higher vocational education is an important part of China’s modern education system, and it has a relationship that promotes and restricts each other’s regional economic development [4]. Regional economy plays a decisive role in the development of higher vocational education and restricts the scale, speed and quality of higher vocational education development to a certain extent. At the same time, vigorously develop higher vocational education and realize its positive interaction with economic and social development. It is directly related to the healthy development of higher vocational education and the implementation of the strategy of strengthening the country by talents, and to the overall situation of regional economic and social development. In the context of the country’s vigorous promotion of the development of higher vocational education, studying the game relationship between the two has far-reaching practical significance. At present, the regional economy is developing rapidly, the industrial structure adjustment and upgrading cover a wide range and great efforts, but the development of vocational education is relatively weak, and there are many problems, such as the inconsistency of professional settings and regional economic structure, the employment structure is not compatible with regional economic development, The traditional teaching model is inconsistent with social practice requirements, the layout structure of economic development is not compatible with the layout structure of higher vocational education, etc., which is not conducive to the development of the regional economy in the region. Since higher vocational education directly serves the industry or local economic development, how to determine the direction and scale of running a higher vocational education based on social needs, how to set majors and determine the needs around the city’s economic development needs and strategic development direction Talent training objectives and models have become the fundamental tasks for the development of higher vocational education in this region. It is particularly important and urgent to study and discuss the relationship between regional vocational education and economic development. The game model has been widely used in many fields due to its many advantages [5, 6]. However, the traditional game model for the development of the relationship between regional education and economy has the disadvantage of slow analysis speed. Association rule mining algorithms can mine associations between different data from big data [7–10]. Based on the above analysis, a game model of regional education and economic development based on association rule mining algorithm is proposed.

154

S. Wang

2 Research on the Game Model of Regional Education and Economic Development Based on Association Rule Mining Algorithm Aiming at the relationship between regional education and economic development, a regional education and economic development game model based on association rule mining algorithm is used to study it. The overall relationship between regional education and economic development is shown in Fig. 1.

Regional economic development

Provide intellectu al support

Provide economic foundatio n

Regional education

Fig. 1. The overall relationship between regional education and economic development

Using the game model of regional education and economic development based on association rule mining algorithm, the specific steps to study the relationship between regional education and economic development are described below. 2.1 Regional Education and Economic Data Collection and Processing First, the relevant data is retrieved from the regional education and economic development database to realize the collection of regional education and economic data. After that, make preliminary processing on the above data. Select the regional education and economic development index system: an index that reflects the development level of higher education. The development level of higher education in a region can be reflected by several indicators. The differences in these indicators can be used to compare the overall level of higher education development in various regions. Eleven indicators reflecting the development level of higher education were selected, namely: enrollment of junior college students per 10,000 people, junior college graduates per 10,000 people, readings

Game Model of Regional Education and Economic Development

155

per 10,000 master students, readings per 10,000 doctoral students, Number of universities, number of students burdened by each university teacher, proportion of employed population in higher education, per capita higher education funding, number of college graduates and above per 100,000 people, average number of students per university and income from major foreign institutions. Select indicators that reflect the level of economic development: Similarly, select 10 indicators that reflect the level of economic development: GDP per capita, GDP growth rate, per capita disposable income of urban residents, per capita net income of rural residents, increase in secondary industry The proportion of value in regional GDP (referred to as the value added of the secondary industry), the proportion of value added in the tertiary industry, the urbanization rate, the consumption rate, the proportion of total imports and exports in various regions in the total national imports and exports, and the per capita regional fiscal revenue. When calculating the degree of order of parameters, for time-varying data, such as GDP, GDP per capita, etc., the maximum and minimum values are bounded, and the country has indicators for setting standards, such as the teacher-student ratio (1:14) and other related Indicators are actually calculated in accordance with nationally prescribed standards. 2.2 Defining the Subject of Regional Education and Economic Development Establish regional economic development and regional education relations. Because the regional government plays a very important role in regional education and economic development, consider it into the relationship between the two, and speculate based on the game, through scientific calculation and evaluation, to study regional college talent training and regional economic development Trend. Assume the following: regional government G, regional economy E, regional university U, now E regional economy is guaranteed and supported by regional government G to apply for a talent program of 10,000 to regional university U, to carry out the construction of local talent industry innovation incubation center, university the return on investment is r. Because of the guarantee and support of the regional government G, the university U believes that the regional economy E has the ability to develop, and that the university U is also a local university and has a close relationship with the regional government G. Therefore, the regional economic development has obtained talents and technology with higher efficiency. Stand by. According to the objective conditions of regional government G between regional economy E and regional university U, the following assumptions are made: the regional government has two kinds of behaviors: contract compliance and breach of contract. The probability of compliance is PG (0 ≤ PG ≤ 1) and the probability of default is 1 − PG . Therefore, at PG = 1 o’clock, the regional government was very committed, and three years after the construction and operation of the talent industry innovation incubation center, the regional government actively urged the regional economic industry to increase the number of talents in regional universities. At PG = 0 o’clock, the regional government completely defaulted, and the regional finances did not have the ability to support the regional economy E. They could only delay it, and the subsequent government continued to improve. Regional economy E has two types of behaviors, honesty and dishonesty. The probability of honesty is 1 − PL and the probability of dishonesty is 2 PL (0 ≤ PL ≤ 1).

156

S. Wang

When PL = 1, the regional economy loses trust and the regional economy resorts to the regional government. At PL = 0 o’clock, regional economic integrity, with the support of regional governments and the help of regional finances, absorb talents from regional universities U on time. As a support and guarantor, the regional government G has financial and policy pressures. If regional economic and industrial innovation incubation is successful, regional finance will have income BL . The support and guarantee cost of the regional government G is CL (financial fund guarantee, state-owned land support, government guarantee, etc.; after the regional economic industry innovation incubation E successfully receives talent and scientific research support from the regional university U, it will be further absorbed at maturity, and the benefit will be L(r − i). When the regional economic industry loses trust, the gain is L(1 + r), and its reputation is lost ML . Through the above, the establishment of the relationship between regional education and economic development is completed. 2.3 Association Rule Mining After processing the data and making relevant assumptions, the Apriori algorithm is used to perform association rule mining. Let the number of fields (attributes) of the regional education and economic data to be analyzed be m and the number of data table records be n. Discretize the data into the form shown in Table 1: Table 1. Discretized data table TID A B

C



1

1

0

1



2

1

0

0



3

0

1

1



4

1

0

0





… … … …

Next, generate a set I of m attributes, such as [A, B, C]; find all subsets of Li (except for the empty set and the full set, i = 1,2, …, 2  −2), Such as [B, C], [C], [A, B], [B], [A, C], [A]; and define a counter Ci for each subset. After that, scan the database and obtain a record set R based on 0 (indicating no inclusion) or 1 (indicating inclusion) for each record. The record set corresponding to the first record in the regional education and economic data is [A, C]. The record set corresponding to the second record is [A], and the record set corresponding to the third record is [B, C] … After that, we use the subset Li of I to match each record set R. If Li ⊆ R, then Ci increases by 1 …

Game Model of Regional Education and Economic Development

157

After the database scan is over, Ci * 100/n can be calculated to obtain the support degree of each subset Li. At this time, you can filter based on the minimum support degree to remove the non-compliant subsets. Combine the remaining subsets in pairs, as shown in Table 2. Then calculate the confidence based on the counts of Li ∪ Lj and Li, and then output the association rules that meet the requirements according to the minimum confidence. Table 2. Association rule generation schematic table [A] [B] [C] [A, B] [A, C] [B, C] [A, B, C] [A]

×

[B]



[C]





[A, B]

×

×

[A, C]

×

→ ×

[B, C]



×

×

×

×

[A, B, C] ×

→ → ×

×



×





×

×



×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

×

→ ×

The set of attributes I is [S_1, S_ 2, W_1, W_2, E-1, E_2, E-3]; find all valid subsets Li of I. In theory, there are 126 subsets except for the empty set and the full set, but only 35 are really effective, and some of the subsets are shown in Table 3. Scan the database. For each record, obtain a record set R based on 0 (indicating that it does not contain) or 1 (indicating that it contains). For example, the record set corresponding to the first record in the database in the table above is {S_1, W_1, E_1}, A subset of I is used to match each record set R. If Li ⊆ R, Ci is increased by one. For example, the subsets {S_1, W_l, E_1}, {W−1}, {S−1, E_1}, etc. are all subsets of the record set, so their corresponding counters are incremented by 1. After the database scan is over, Ci * 100/n can be used to obtain the support of each subset Li. At this time, you can filter based on the minimum support (assuming 50%) to remove the subsets that do not meet the requirements. As shown in Table 3, the subsets {S_2, W_2, E_1}, {W_2, E_1}, {W_1, E_1}, etc. do not meet the requirements, while the subsets {S_l, W_1}, {W_1}, {S_1, W_2}, etc. meet the requirements … In order to make the attribute classification more explicit, it is divided in the form of a matrix. Let attribute U = {u1 , u2 , u3 , u4 , u5 } be the set of divided attribute objects. If the corresponding classification matrix is: ⎡

⎤ 10100 R = ⎣0 1 0 0 1⎦ 00010

(1)

158

S. Wang Table 3. Attribute subset and count Serial number SubsetLi

CounterCi

1

{S_2, W_2, E_1}

0

2

{S_1, W_2, E_1}

0

3

{W_2, E_1}

0

4

{S_2, W_1, E_1}

1

5

{S_1, W_1, E_1}

5

6

{W_1, E_1}

0

{S_2, E_1}

0

{S_1, E_1}

0

{S_1}

1

{S_2, E_2}

5

{S_1, W_2}

10

{W_2}

6

{S_2, W_1}

4

{S_1, W_1}

10

{W_1}

14





The corresponding classification result is {u1 , u3 }, {u2 , u5 }, {u4 }. For example, the corresponding classification matrix is: ⎡ ⎤ 10010 R = ⎣1 0 0 0 1⎦ (2) 00100 The corresponding classification results are {u2 , u4 }, {u1 , u5 }, and {u3 }. On this basis, fuzzy concept can also be used for fuzzy classification, that is, the object u in the classified object set U is considered; it belongs to a certain class with a certain degree of membership, not simply “1” (belonging to) or “0” (Not belong to), is a number between 0 and 1 to indicate the extent to which an object belongs to a certain class. Therefore, each class is considered to be a fuzzy subset of the object set U, so each such The classification corresponding to the classification result is a fuzzy matrix R: ⎤ ⎡ r11 r12 · · · r1n ⎢ r21 r22 · · · r2n ⎥ ⎥ ⎢ (3) R=⎢ . . . . ⎥ ⎣ .. .. .. .. ⎦ rc1 rc2 · · · rcn

Each fuzzy classification of the object set U is classified into c class corresponds to a fuzzy matrix R that meets certain conditions; conversely, any fuzzy matrix R that meets a

Game Model of Regional Education and Economic Development

159

certain condition also corresponds to the fuzzy classification of object set U is classified into c class. Finding the best fuzzy classification matrix R under a certain condition, the fuzzy classification corresponding to R is the best fuzzy classification of the object set U under this condition. Combining the subsets that meet the requirements (but must meet Li (Lj = Q), and there is no relevant attribute in Li ∪ Lj) to generate the association rule Li → Lj. As shown in Table 3, the support levels of the subsets {S_1, W_2} and {W_1} meet the requirements, but the rule “{S−1, W_2} → {W−1}” cannot be generated because W_2 and W_1 are related The attributes of the subsets {S_ 1} and {W_1} meet the requirements, both 14 * 100/20 = 70%, and these two attributes are irrelevant, and the rules can be generated after crossing “{S_1} → {W_1}” and {W−1} → {S_1}; since the subset {S_1, W_1} (that is, the union of the two subsets {S_1} and {W_1}) is counted at 10, this time is also possible Calculate the confidence of these two association rules. Through confidence, use association rules to fully mine regional education and economic data. 2.4 The Game of the Relationship Between Regional Education and Economic Development Based on the Degree of Coupling Function Use the association rule mining algorithm to fully mine regional education and economic development data. According to the above-mentioned game scenario simulations and hypothetical conditions, the utility of regional government compliance is 1 UG1 , and the utility of breach is UG2 , then: UG1 = PL (−CL ) + (1 − PL )(BL − CL )

(4)

UG2 = PL + (1 − PL )BL

(5)

Because the regional government is a regional administrative organ, with functions such as policy and social leadership, and a strong organizational credit foundation and authority structure foundation, when the regional government supports and guarantees the regional economy, it will only keep the contract when UG1 > UG2 Can be solved: F −H >

C PL

(6)

In formula (6), in order to obtain PL , it is assumed that the trustworthy regional government economy allocates a certain income HL to the regional government, and gains L(1 + r) after receiving the support of the talents of the regional colleges and universities, so the reputation is lost ML . The benefit of the regional economy after integrity is L(r − i); if the regional economy is not supported by talents, the opportunity loss is L(r − i); the U talents of regional universities support the regional economy that does not return on time, and loses the benefit L(1 + i); the talents support the regional economy that returns on time, The return is Li . If regional universities do not provide talent support as agreed, the opportunity loss is Li . The probability of talent output from regional universities is PB , the probability of rejection is (1 − PB ); the probability of regional

160

S. Wang

economic mistrust is PL , and the probability of integrity is (1 − PL ), so that the game payment matrix between regional economy E and regional university U can be obtained as shown in the Table 4 shown. Table 4. Game payment matrix between regional economy and regional universities Regional economy L Broken promise PL Regional colleges U

Talent export PB [−L(1 + i), L(1 + r) − ML − HL ]

Honesty (1 − PL )  −Li , L(r − i)

Reject output (1 − PB )

[−Li , −L(r − i)]



−Li ,



−L(r − i)

Coupling degree A is obtained through the game between regional economic incubation center E and regional university U. The degree of coupling refers to the degree of influence between the two. Here, the degree of coupling between regional education and the economic system can be obtained by calculation, and A ∈ [0, 1]. The larger the value, the higher the coupling level, and the smaller the value, the more uncoordinated. When the value of A is zero, the coupling level is the lowest, and when it is 1, the coupling level is the highest. Set the sequence parameter of the higher education subsystem to V1 and the sequence parameter of the regional economic subsystem to V2 , the specific formula is as follows: √ 2 V1 · V2 A= (7) V1 + V2 In this way, it can be concluded that the maximum utility function of the regional economy is UL : UL = F − A >

C(M + H − I − 1) 2(r − i)

(8)

At this point, the establishment of a regional education and economic development game model based on association rule mining algorithms is completed.

3 Experiment Compare the proposed regional education and economic development game model based on association rule mining algorithm with the traditional regional education and economic development game model to verify whether the proposed regional education and economic development game model based on association rule mining algorithm can be more Get analysis results quickly.

Game Model of Regional Education and Economic Development

161

3.1 Experiment Procedure Taking Jiangxi Province as an example, a game model of regional education and economic development based on the association rule mining algorithm is used to study the relationship between education and economic development. Taking the data from 2006 to 2019 as the analysis object, the number of students in the vocational colleges in the region, and the per capita GDP (GDP per capita) were used as indicators to measure the state of education and economic development in the region. Based on the results, explore the relationship between the two. The situation of the two is shown in Table 5. Table 5. Students and GDP per capita of higher vocational colleges in a region from 2006 to 2019 Years Student (person) GDP per capita (yuan/person) 2006 432296

23601

2007 500981

27506

2008 805145

32146

2009 574976

35781

2010 597846

41125

2011 557140

47225

2012 534679

51478

2013 548978

56784

2014 614785

60147

2015 594716

64152

2016 584989

67845

2017 601245

70129

2018 612348

73248

2019 597856

78453

According to the above data, the relationship between the two is analyzed through the proposed game model. For two different game models, the analysis speed is compared. 3.2 Analysis of Results Using the proposed regional education and economic development game model based on the association rule mining algorithm and the traditional regional education and economic development game model to analyze the relationship between regional education and economic development, and compare the analysis speed of the analysis results, such as Shown in Fig. 2. As can be seen from the figure, from 2015 to 2019 as an example, using the traditional regional education and economic development game model, the time required to obtain the analysis results is more than 2 s; the proposed regional education and economics

162

S. Wang Proposed game model Traditional game model

Analysis time / s

2.5

2.0

1.5

1.0

0.5

2015

2016

2017

2018

2019

Year / year

Fig. 2. Analyze speed comparison results

based on association rule mining algorithm is used The development of the game model, based on the association rule mining algorithm, analyzes the data of regional education and economic development, speeding up the analysis speed, and the required analysis time is about 0.5~0.6 s. Through comparison, it is found that the proposed regional education and economic development game model based on association rule mining algorithm can obtain the analysis results more quickly. 1.2 1.0 0.8 0.6 0.4 0.2 0

2005

2006 2007

2008

2009

2010

2011

2012

2013

2014

2015

2016 2017

Higher education development level

Regional economic development level

Coupling C

Coupling coordination D

Fig. 3. Coupling analysis

Figure 3 shows that the level of development of the regional economic higher education system in Jiangxi Province from 2005 to 2017 has increased year by year. The degree of coupling is also maintained at a high level of about 0.99, but the level of coupling coordination of the system is not satisfactory. The degree of coupling coordination between 2005 and 2007 was only about 0.2, which is in a state of moderate imbalance. This is because in recent years not only the level of higher education in our province is low, but also the level of regional economy is very low. In the three years from 2008 to 2010, the degree of coupling and coordination between higher education and the regional economy in our province was in a state of slight imbalance, and in the period of 2011 to 2012, it was on the verge of imbalance. Beginning in 13 years, the transition from

Game Model of Regional Education and Economic Development

163

imminent imbalance to a state of barely coordination. From the data in the appendix, it can be seen that the investment in science and technology projects in Jiangxi Province in 2013 was ten times that in 12 years, and the research results of that year have also been greatly improved. This is likely to be an important reason for the rapid development of coupling and coordination. One. The degree of coupling and coordination of the system in our province began to reach the state of primary coordination in 2016. The degree of coupling coordination is increasing year by year, and from the perspective of the value gap between adjacent years, the value difference between adjacent years in 05–17 is not greater than 0.1, and both are in the range of [0.01, 0.07].

4 Policy Suggestions for Coordinated Development of Higher Education and Regional Economy According to the coupling analysis results in Fig. 3, the coordinated development of higher education and regional economy needs to follow certain policies. First, we should increase funding for higher education to promote the transformation of teaching and research results; second, we should adjust the discipline and professional structure, and local universities should actively adapt to the needs of economic and social development in Jiangxi Province, meet the needs of knowledge innovation, scientific and technological progress, and education development, and plan rationally. Professional layout and development, dynamic adjustment of majors, to make the professional structure more reasonable, to optimize the professional layout, to provide space for the sustainable development of the profession, and to strive to achieve the structural balance between benign talent training and the demand for talents and the industrial characteristics of a healthy interaction. In the end, special industries in Jiangxi Province should be developed, with characteristic industrial chains as a link, to form clusters with certain competitive advantages, and promote local economic development.

5 Concluding Remarks Aiming at the disadvantages of the slow analysis speed of the traditional regional education and economic development game model, a regional education and economic development game model based on association rule mining algorithm was proposed. Through comparative experiments, compared with the traditional regional education and economic development game model, the experimental results show that the proposed regional education and economic development game model based on the association rule mining algorithm can obtain the analysis results more quickly. Using this game model, it is possible to better analyze the relationship between regional education and economic development, which is conducive to the common development of education and economy, and to make the professional structure more reasonable, achieve a balance between talent training and talent demand, and form a certain competition Advantageous clusters promote local economic development. Fund projects. The 13th Five Year Plan of Jiangxi Education Science “Research on the Coordinated Development of Regional Higher Education and Economy” (17ZD065).

164

S. Wang

References 1. Wang, Z.: Research on the collaborative development of higher vocational education and regional economy from the perspective of triple helix model. J. Zhejiang Bus. Technol. Inst. 18(1), 78–81 (2019) 2. Li, J., He, S., Zhang, Q.: The research on the relationship between development of higher education and economic growth in Jilin Province: based on the grey relation entropy model. Jilin Normal Univ. J. (Human. Soc. Sci. Ed.) 46(1), 89–96 (2018) 3. Tang, Y.: System innovation of coordinating urban and rural vocational education and regional economic linkage development. J. Heilongjiang Coll. Educ. 38(9), 57–59 (2019) 4. Zhang, H., Huang, J., Cui, Y.: Game models and contract of green supply chain considering fairness preference and government subsidies. Ind. Technol. Econ. 37(1), 111–121 (2018) 5. Li, X., Liu, C., Zhong, J.: User-centered demand response game model and algorithm research. J. Nanjing Inst. Technol. (Nat. Sci. Ed.) 16(2), 16–21 (2018) 6. Bi, H.: Research on the relationship between the subjects of marathon entry based on the bilateral matching game model. J. Jilin Sport Univ. 35(3), 26–31 (2019) 7. Xiao, W., Hu, J., Zhou, X.: Parallel association rules mining algorithm based on MapReduce: a survey. Appl. Res. Comput. 35(1), 13–23 (2018) 8. Zeng, Z., Gong, Q., Zhang, J.: Improved association rule mining algorithm: MIFP-apriori algorithm. Sci. Technol. Eng. 19(16), 216–220 (2019) 9. Lu, X., Wang, X.: Educational administration data mining of association rules based on domain association redundancy. Comput. Sci. 46(z1), 427–430,435 (2019) 10. Zhao, J., Liu, Y., Liu, N., et al.: Association rules of monitoring and early warning by using landslides FRPFP model—case study of Jiangjin-Fengjie reach in Three Gorges Reservoir area. Chin. J. Geotech. Eng. 41(3), 492–500 (2019)

Automatic Detection of Image Features in Basketball Shooting Teaching Based on Artificial Intelligence Sha Yu1(B) and Jing Liu2 1 Sports Department of Shaanxi, University of Traditional Chinese Medicine,

Xianyang 712046, China [email protected] 2 School of Mathematics and Statistics, Nanyang Institute of Technology, Nanyang 473000, China

Abstract. In order to weaken the limitation between the adjacent nodes of shooting action teaching image and fully expose the necessary image features, an automatic detection method of shooting action teaching image features based on artificial intelligence is proposed. Starting from the decomposition of shooting teaching action, the practical attribute of shooting teaching image is defined, and then the standardized image detection environment based on artificial intelligence is built according to the action essentials of consistent shooting. On this basis, through the way of extracting the characteristics of teaching image, calculating the characteristic scale value of shooting action, combining with the feature description principle of automatic detection, the design of automatic detection method of shooting action teaching image characteristics is completed. The experimental results show that, compared with the traditional feature method, the UTI traction coefficient between the adjacent nodes of the image increases to 0.33, while the matching time required for detection decreases to 17.5 mm, so as to fully expose the necessary features of the shooting action teaching image. Keywords: Artificial intelligence · Shooting action · Image features · Automatic detection · Action decomposition · Characteristic scale · Detection characteristics · Traction coefficient

1 Introduction Artificial intelligence stands for artificial intelligence. It is an emerging technical science to research and develop theories, methods, techniques and application systems that simulate, extend and extend human intelligence. Artificial intelligence is a branch of computer science. It tries to understand the nature of intelligence and produce a new kind of intelligent machine that can react in a way similar to human intelligence. Research in this field includes robotics, language recognition, image recognition, natural language processing and expert systems [1]. Since the birth of artificial intelligence, its theory and technology © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 165–175, 2020. https://doi.org/10.1007/978-3-030-63955-6_15

166

S. Yu and J. Liu

have become increasingly mature, and its application field has also been expanding. It is conceivable that the technological products brought by artificial intelligence in the future will be the “containers” of human intelligence. Artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is not human intelligence, but it can think like a human, and may surpass it. Feature detection is a concept in computer vision and image processing. It refers to the use of computer to extract image information and determine whether each image point belongs to an image feature. The result of feature detection is to divide the points on the image into different subsets, which are often isolated points, continuous curves or continuous areas. So far, there is no universal and precise definition of features. The exact definition of a feature is often determined by the problem or application type. Feature is an “interesting” part of a digital image, which is the starting point of many computer image analysis algorithms [2]. Therefore, the success of an algorithm is often determined by the characteristics it uses and defines. Therefore, the most important feature of feature detection is “repeatability”: the features extracted from different images of the same scene should be the same. With the popularization of basketball teaching methods, how to analyze the characteristic behavior of the formed image according to the existing shooting action so as to realize the automatic detection has become the main research direction in related research fields. In order to solve the above problems, the traditional feature method checks whether each teaching action pixel represents a feature through primary operation. If it is part of the extended image algorithm, then this pixel point is only used to check the feature area of the image. However, the actual detection accuracy of this method is very low, and it is difficult to meet the increasingly strict shooting action teaching task. Based on this, the concept of artificial intelligence is introduced to design a new type of shooting action teaching image feature automatic detection method, and through the edge pixel point statistics and other methods, the detection indicators are debugged to the best application state.

2 Standardized Image Detection Environment Based on Artificial Intelligence The construction of standardized image detection environment is the basic link of the realization of the automatic detection method of shooting action teaching image characteristics. Under the support of artificial intelligence theory, according to the operation process of shooting teaching action decomposition, shooting action teaching image practical definition and consistent shooting action essentials analysis, the specific construction and processing methods are as follows. 2.1 Shooting Teaching Action Breakdown Artificial intelligence is a subject that studies how to make computer simulate some thinking process and intelligent behavior (such as shooting action) of human beings. It mainly includes the principle of realizing intelligence by computer, manufacturing computer similar to human brain intelligence, and making computer realize higher-level

Automatic Detection of Image Features

167

application. Artificial intelligence will involve computer science, psychology, philosophy and linguistics. It can be said that it is almost all disciplines of natural science and social science, and its scope is far beyond the scope of computer science. The relationship between artificial intelligence and thinking science is the relationship between practice and theory. Artificial intelligence is at the technical application level of thinking science, and is an application branch of it [3]. From the point of view of thinking, AI is not only limited to the logical thinking of shooting action, but also considers the image thinking and inspiration thinking of teachers to promote the breakthrough development of AI. Mathematics is often considered as the basic science of image feature detection. Mathematics has also entered the field of language and thinking. The subject of AI must also borrow mathematical tools to realize the teaching of shooting action Directional perception of image features (Fig. 1).

Fig. 1. Breakdown of shooting teaching action based on Artificial Intelligence

2.2 Practical Definition of Shooting Action Teaching Image The mathematical basis of AI is “Statistics”, “information theory” and “Cybernetics”, including other non mathematical subjects. This kind of shooting teaching needs a strong dependence on “experience”. The computer needs to acquire knowledge and learning strategies from the experience of solving a class of problems. When encountering similar image feature decomposition problems, it uses the experience knowledge to solve problems and accumulate new experience, just like the action of committing a project that has been accumulated for many times. Such a processing method is called “continuous learning” in artificial intelligence [4, 5]. But in addition to learning from experience, teachers can also create, that is, “jumping learning”. In some cases, this is called teaching “inspiration” or “Epiphany”. For a long time, the most difficult thing in the field of artificial intelligence is “Epiphany”. Or strictly speaking, it is difficult for teachers to learn “qualitative change of actions that do not depend on quantitative change” in learning and “practice”, and it is difficult to go directly from one “quality” to another “quality”, or from one “concept” to another “concept”. Because of this, the “shooting teaching practice” here is not the same as the practice of artificial intelligence image. Suppose q represents the statistical coefficient of shooting action teaching image based on the field of artificial intelligence, y1 represents the upper scale range of image features, and y2 represents the lower scale range of image features. Combining the above

168

S. Yu and J. Liu

physical quantities, we can define the practical characteristic indexes of shooting action teaching images as follows: y1 λ= y2

q2 |r1 × r2 | u¯ 2

(1)

Among them, r1 represents the continuous learning coefficient in shooting teaching, r2 represents the practical learning coefficient in shooting teaching, and u¯ represents the average counting condition of characteristic pixels in the image. 2.3 Key Points of Continuous Shooting There are two aspects of shooting technique, one is the body posture when shooting, the other is holding the ball. When shooting in situ, open the front and back of your feet naturally, bend your knees slightly, lean your upper body forward slightly, and place your weight between your feet. It is easy to focus on the shooting force, but also conducive to change other movements. In the process of catching and jumping shots, dribbling and stopping shots or shooting shots during moving, the step catch and take-off should not only be connected, but also be braked quickly, so that the center of gravity of the body can be moved to the center of the supporting surface as soon as possible, so as to ensure the vertical take-off. The correct body posture can ensure that the movement of the body’s center of gravity is consistent with the direction of shooting, and can maintain the body balance. Controlling body balance is the basic condition to ensure the accurate direction of the ball [5]. When shooting, no matter with one hand or two hands, when holding the ball, the five fingers should open naturally and the palm should be free. Touch the ball with the finger root and above the finger root to increase the contact area of the ball, so as to maintain the stability of the ball and control the shooting direction of the ball. For the teaching implementers, the shooting skills can be summed up in two directions: chest shooting with both hands and shoulder shooting with both hands. 2.3.1 Two Hands Chest Shot Although the shooting point is low, the stability is good before the shot, the strength of the shot is large, it is easy to combine with passing and breakthrough, and it is mostly used for long-distance shooting. Movement method: holding the ball with both hands is basically the same as passing the ball in front of the chest with both hands. Drop your elbows naturally, place the ball in front of your chest and look at the aiming point. Open the front and back or left and right feet, slightly bend the knees, and place the center of gravity between the feet (As shown in Fig. 2). 2.3.2 Shoot with Both Hands and Shoulders This kind of shooting points high, not easy to cover, easy to combine with the head pass. But the center of gravity is high, the speed of shooting is slow, and it is easy to shake the center of gravity of the defensive side. Hold the ball on the head with both hands,

Automatic Detection of Image Features

169

Fig. 2. Analysis of the characteristics of the chest shooting with both hands

bend the elbow naturally, open the feet obliquely, bend the knees slightly, and place the center of gravity between the feet. When shooting, the two arms are extended forward and upward along with the leg pedaling, the two wrists are turned outward at the same time, the left hand is pressed down, and the right thumb is used to control the direction of the ball. There are index finger and ring finger at the moment of release. After the ball is released, the lower body should be relaxed.

3 The Automatic Detection Method of Shooting Action Teaching Image Characteristics Based on the theory of artificial intelligence, according to the processing flow of teaching image feature extraction, shooting feature scale calculation and automatic detection feature description, the successful application of shooting action teaching image feature automatic detection method is completed. 3.1 Feature Extraction of Teaching Image The feature nodes of each layer in the shooting action are taken as the initial image to construct the automatic detection scale space, and the sub group scale space including two inner layers and two middle layers is constructed for it, and the image of this layer is taken as the first inner layer image in each sub group. In each subgroup, each action middle layer is distributed between two adjacent inner layers. Each inner image is obtained by 0.5 times lower sampling of the upper inner image. The first middle layer is obtained by 1.5 times down sampling of the first inner layer image, and the other middle layer is obtained by 0.5 times down sampling of the previous middle layer [6]. If the coverage between the scale nodes of the shooting action teaching image is not considered, the detection characteristic value of the feature to be extracted can be expressed as the parameter space as shown below.

170

S. Yu and J. Liu Table 1. Feature extraction value set of shooting action teaching image

Feature extraction value

Action layer

Teaching layer

The image layer

The data layer

Parameter layer

1

0.38

20

0

0.829

0

2

0.42

35

0

1.513

121

3

0.45

37

0

2.391

113

4

0.51

123

113

3.791

96

5

0.59

37

20

3.453

89

6

0.66

44

34

4.326

115

7

0.73

51

21

3.187

128

According to Table 1, when the shooting angle, rotation direction, angle of view, image features and noise remain unchanged, the necessary execution data always has strong robustness, and at the same time, it can achieve some actual image node matching operations that the initial behavior cannot complete. And this improved method between the numerical value and the numerical value can fundamentally alleviate the coverage between adjacent node layers, so that the image features to be detected are fully exposed. 3.2 Calculation of Shooting Characteristic Scale In order to speed up the detection accuracy of teaching image features, firstly, according to the characteristic scale of the input shooting action, the appropriate block processing is carried out. Secondly, multi threads are used to construct new scale space in block image, and necessary numerical feature points are extracted. According to the matching requirements, select a certain number of characteristic nodes in each block area to organize, build the set of characteristic points, and then use multi threads to build the set of characteristic vectors. Then, the feature vector set is processed by two-way matching and eliminating mismatches [7, 8]. Finally, according to whether the number of matching points is greater than the set detection value, if it is greater than, the matching is successful. Otherwise, rebuild the set of feature points and repeat the above process. Let w¯ represent the average value of scale space in the image vector set of shooting action teaching, and Ω represent the value space of image vector set. The joint formula (1) can express the calculation result of shooting characteristic scale as follows.     I ↑   2  λ(U1 − U0 )   I ↓ (2) E =  I↑   w¯ · ||d + d  || I↓

Among them, I ↑ represents the maximum behavior authority value of the characteristic node scale of the shooting action teaching image, I ↓ represents the minimum behavior

Automatic Detection of Image Features

171

authority value of the characteristic node scale of the shooting action teaching image, U1 represents the characteristic image node parameter, U0 parameter value supplementary description condition, d represents the characteristic image vector value, d  represents the vector value supplementary description condition. 3.3 Automatic Detection Feature Description In the field of artificial intelligence processing, the automatic detection of scale invariant features is an important attribute. Traditional feature detection algorithms are based on linear Gaussian pyramid to extract feature points. For example, feature method uses the method of building the frame structure of Gaussian differential scale space, surf algorithm uses the box filter method of approximate Gaussian differential. These methods use Gaussian blur to construct the image Gaussian pyramid, but Gaussian blur not only does not retain the boundary information of the object, but also smoothes the details and noise to the same degree on all scales, thus losing the positioning accuracy [9]. In order to deal with the image data of shooting movement teaching, the method of feature detection and description in nonlinear scale space has been proposed. However, the traditional method is based on the forward Euler method to solve the nonlinear diffusion equation, which has the advantages of long convergence step, long time and high computational complexity, In order to solve the above problems, an automatic feature detection method based on non-linear scale space is introduced and implemented, which can detect and match shooting action images with the support of multi-scale nodes. Let g1 represent the scalar characteristic value in the shooting action teaching image, and g2 represent the vector characteristic value in the shooting action teaching image. The simultaneous formula (2) can define the automatic detection characteristic description as. √ l 2 − 1 g1 g2 (3) K = exp(− 2 ) · βf 2 E In the above formula, l represents the comprehensive behavior change of the shooting action teaching image characteristic value in unit time, β represents the vector index coefficient of the automatic detection behavior, and f represents the vector linear error condition carried by the image characteristic node. So far, we have completed all previous calculations and debugging of numerical error results, and completed the construction of automatic detection method of shooting action teaching image features with the support of necessary AI behavior instructions.

4 Practical Application Detection In order to highlight the practical difference between the automatic feature detection method based on artificial intelligence and the traditional Feature method, a comparison experiment is designed as follows. In the course of shooting teaching, we select the necessary movement features as the reference object, and take the host computer of the new automatic detection method and the traditional Feature method as the parameter recording equipment of the experiment group and the control group. In the same detection environment, we study the specific changes of the image feature indexes of the experiment group and the control group.

172

S. Yu and J. Liu

4.1 Detection Application Environment In order to ensure the absolute fairness of the experimental results, the ball selection processing in the whole experimental process is completed by the artificial intelligence robot. The relevant teaching participants are only as the detection and verification personnel to supervise and inspect the ball selection operation of the robot (Figs. 3 and 4).

Fig. 3. Artificial intelligent ball selection

Fig. 4. Characteristic image of shooting action teaching

4.2 UTI Traction Coefficient The UTI traction coefficient directly affects the final detection results of shooting action teaching image features. Generally, the greater the value level of the former, the higher the accuracy of the latter, and vice versa. The following table reflects the specific comparison of UTI traction coefficient between the experimental group and the control group. Table 2 shows that as the experiment time goes on, the UTI traction coefficients of the experimental group keep increasing and decreasing alternately, the global maximum is 0.34, the minimum is 0.07, and the difference between them is 0.27.

Automatic Detection of Image Features

173

Table 2. UTI traction coefficient of experimental group The experimental time/(min)

UTI traction coefficient

The average 0.20

5

0.07

10

0.28

15

0.09

20

0.31

25

0.09

30

0.30

35

0.07

40

0.34

45

0.08

50

0.33

55

0.07

60

0.31

Table 3 shows that the UTI traction coefficients of the control group increased steadily in the early stage of the experiment, and kept stable for 20 min after reaching the extreme value. The global maximum was only 0.17, which was 0.17 lower than the extreme value of 0.34. In conclusion, with the application of automatic feature detection method Table 3. UTI traction coefficient of experimental group The experimental time/(min)

UTI traction coefficient

The average 0.13

5

0.11

10

0.13

15

0.15

20

0.17

25

0.17

30

0.17

35

0.17

40

0.17

45

0.16

50

0.10

55

0.04

60

0.02

174

S. Yu and J. Liu

based on artificial intelligence, the UTI traction coefficient increases obviously, which promotes the accuracy of automatic feature detection to some extent. 4.3 Detection of Matching Time In general, the longer the time of feature matching is, the worse the precision of the detection results is, and the worse the quality of the detection results is. The following figure reflects the specific changes of the matching time in the experimental group and control group under the same experimental environment.

Fig. 5. Experimental group detection matching time

Fig. 6. Matching time in control group

Compared with Figs. 5 and 6, the peak value of the matching time of the experimental group was only about 20 mm, and the limit value did not have the ability of periodic existence; the peak value of the matching time of the control group was more than 25 mm, much higher than that of the experimental group, and the limit value had the ability of periodic existence. In conclusion, with the application of automatic feature detection method based on artificial intelligence, the matching time of feature parameters decreases obviously, and the accuracy of automatic feature detection is improved properly.

Automatic Detection of Image Features

175

5 Conclusion Along with the application of artificial intelligence technology, the automatic feature detection method of shot motion teaching image not only redefines the concept of attributes of relevant indexes, but also analyzes the algorithm based on the conditions of detection and processing from the point of view of coherence, which not only constrains the vector scale value of shot feature, but also reduces the influence of unnecessary error value on the final detection precision. From the point of view of practical application, the matching time has been effectively controlled, and the UTI traction coefficient has been moderately improved, which solves the problem of inadequate exposure of image features left over by the traditional Feature method.

References 1. Shoufeng, J., Di, F., Reza, M., et al.: An image recognition method for gear fault diagnosis in the manufacturing line of short filament fibres. Insight: Non-destr. Test. Condition Monit. 60(5), 270–275 (2018) 2. Zhang, Z., Bian, J., Han, W., et al.: Cotton moisture stress diagnosis based on canopy temperature characteristics calculated from UAV thermal infrared image. Nongye Gongcheng Xuebao/Trans. Chin. Soc. Agric. Eng. 34(15), 77–84 (2018) 3. Ahmed, T., Sarma, M.: Locality sensitive hashing based space partitioning approach for indexing multidimensional feature vectors of fingerprint image data. IET Image Proc. 12(6), 1056–1064 (2018) 4. Lahouli, I., Karakasis, E., Haelterman, R., et al.: Hot spot method for pedestrian detection using saliency maps, discrete Chebyshev moments and support vector machine. IET Image Proc. 12(7), 1284–1291 (2018) 5. He, X., Li, L., Liu, Y., et al.: A two-stage biomedical event trigger detection method integrating feature selection and word embeddings. IEEE/ACM Trans. Comput. Biol. Bioinf. 15(4), 1325– 1332 (2018) 6. Alghowinem, S., Goecke, R., Wagner, M., et al.: Multimodal depression detection: fusion analysis of paralinguistic, head pose and eye gaze behaviors. IEEE Trans. Affect. Comput. 9(4), 478–490 (2018) 7. Malak, J.S., Zeraati, H., Nayeri, F.S., et al.: Neonatal intensive care decision support systems using artificial intelligence techniques: a systematic review. Artif. Intell. Rev. 21(1), 1–20 (2018) 8. Cao, G., Zhang, H., Zheng, J., et al.: An Outlier Degree Shilling Attack Detection Algorithm Based on Dynamic Feature Selection. Int. J. Softw. Eng. Knowl. Eng. 29(8), 1159–1178 (2019) 9. Chen, G.S., Zheng, Q.Z.: Online chatter detection of the end milling based on wavelet packet transform and support vector machine recursive feature elimination. Int. J. Adv. Manuf. Technol. 95(5), 1–10 (2018)

Design of Higher Education Aided Teaching System Based on Analysis of Economic Coupling Coordination Degree Shan Wang(B) School of Engineering and Management, Pingxiang University, Pingxiang 337000, China [email protected]

Abstract. Aiming at the problem of the unsatisfactory feedback effect of the traditional auxiliary teaching system, a higher education auxiliary teaching system based on the analysis of economic coupling coordination degree was designed. The auxiliary teaching tasks are assigned based on the degree of economic coupling and coordination, and the interactive feedback of the teaching system is realized by setting the distributed filtering method of the auxiliary system to the teaching content. Experimental results show that compared with the two traditionally designed auxiliary teaching systems, the higher education auxiliary teaching system designed this time has better feedback effect and stronger system stability. The system can be applied to higher education auxiliary teaching tasks. Keywords: Economic coupling consistency · Higher education · Auxiliary teaching system

1 Introduction According to the characteristics of network teaching, the auxiliary teaching system strives to break through the limitations of traditional auxiliary teaching systems, and designs and implements a network teaching system based on the analysis of economic coupling and coordination. Analysis of economic coupling coordination degree, setting up educational resource allocation tasks according to the rule mode, and realizing a more targeted and interactive system design. Connect the system to the course resource website, and set up teaching courses, add course-related teaching content, students learn online by browsing the course website, or download course-related materials for selfstudy, submit assignments online, communicate with teachers and classmates online, online Exam and check your results. Teachers can post related teaching content online, view and correct student assignments, reply to student messages, and more. The higher education auxiliary teaching system designed this time provides a platform for communication between students and teachers, students and students, breaks the constraints of time and space, autonomously studies anytime, anywhere, and provides interactive data feedback through strong technical support [1]. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 176–188, 2020. https://doi.org/10.1007/978-3-030-63955-6_16

Design of Higher Education Aided Teaching System

177

2 Hardware Design of Higher Education Auxiliary Teaching System In order to ensure the reliability of the aided teaching system, the hardware of the aided teaching database cluster system is redesigned. In order to ensure fast instruction execution and high execution efficiency, to ensure that data operations are completed in registers, the ARM processor is replaced. The new ARM processor model is ARM-A33. This processor is shown in Fig. 1.

Fig. 1. Processor physical diagram

The processor is connected with the central control module of the system, and the communication mode between the main control equipment and the analysis hardware is established to form a complete system hardware architecture to control the receiving of the access data by the auxiliary teaching system. At the same time, the DSP processor in the original system should be replaced, the processor should be connected to the sensor, and the digital information with the characteristic of volume should be converted into electric signal to find, identify, process and feedback the massive data in the database quickly [2]. By redesigning the hardware of these two systems, the cluster system can complete the effective operation of the hardware under the economic coupling coordination analysis in one instruction cycle, and ensure that there is a gap between the program and the data space, and the query and search data can be accessed simultaneously to realize the distributed general sharing management of the system. The type of DSP processor is J380, the applicable working temperature range is between −30 °C and 60 °C, the signal-to-noise ratio is more than 92 dB with static noise, and it is more than 85 dB without static noise, which conforms to the design and use requirements of the teaching database cluster system.

178

S. Wang

3 Design of Higher Education Auxiliary Teaching System Based on Economic Coupling Coordination Analysis 3.1 Allocation of Auxiliary Teaching Tasks Based on Economic Coupling Coordination In order to ensure that the task assignment of the auxiliary teaching system conforms to the actual work requirements, the database cluster access node is set up to ensure the real-time, synchronization and modifiability of the assigned tasks. The rules for setting the access node are shown in Fig. 2 below [3].

Fig. 2. Node setting rules

From Fig. 2, it can be seen that the large square indicates the central node; the small square indicates the secondary central node; the straight line with the arrow indicates the connection path between the nodes; and the dotted line indicates the degree of association between the data in each node. Assume that the central node is c0 and the secondary central node is cn = {c1 , c2 , . . . , cn }, where n represents the total number of secondary nodes and is a non-zero natural number. Set the processing request parameters of each node. The parameters are shown in Table 1 below [4]. Table 1. Node processing request parameters Node number Weight Processing request parameters c1

w1

F * w1/( w1 + w2 + w3 + …+ wn)

c2

w2

c3

w3





F * w2/( w1 + w2 + w3 + …+ wn) F * w3/( w1 + w2 + w3 + …+ wn) …

cn

wi

F * wn/( w1 + w2 + w3 + …+ wn)

F in the table indicates a rule parameter. According to the request processing parameters obtained in the above table, the CRITIC weighting method and the Mahalanobis

Design of Higher Education Aided Teaching System

179

distance calculation method are combined to set the relative distance between the nodes. The calculation expression of the distance is: ⎧      ⎪ ⎨ d (Xi , γ ) =  xi − γ +  −1 xi − γ − T (1)  ⎪ ⎩ d Y , γ  =  y − γ +  −1 y − γ − T j j j In the formula: d represents the distance between any two adjacent points; Xi represents the node to be calculated at position i; Yj represents the node to be calculated at j closest to i; γ + represents a positive ideal solution; γ − represents negative ideal solution;  represents the covariance matrix;  −1 represents the inverse matrix; T represents the matrix transpose sign. Determine the effective control distance between nodes according to the above formula, and make sure that all information can pass smoothly without affecting the balance of the load [5]. Based on the node position set above, a threshold σ is introduced to divide all the data to be processed in the database into light load nodes and heavy load nodes according to the processing nodes. When the load value is μ > σ , the node is a heavy load node. When μ < σ , the node at this time is a light load node. When performing a teaching query task, all nodes begin to execute query search instructions. Assume that the light and heavy nodes are evenly distributed in the data space, so that the load value of the heavy load node is μ1 , the load value of the light load node is μ2 , and the total number of nodes in the relevant domain is a. The function expression of the average load is: σa + η μ¯ =

n

wi

i=1

a+1

(2)

In the formula: σa represents the load value when the total number of nodes is a; wi represents the data load corresponding to the node at the i node; η represents the standard index of the degree of coordination of economic coupling. In order to ensure that the data selected by the selected nodes to the visiting users are reliable, introduce a control coefficient A that can avoid load migration. Based on this coefficient, calculate the load amount. The calculation expression is: μ =

¯ fa (σa − μ) n η fi

(3)

i=1

In the formula: μ represents the calculation result after the load changes; fi represents the control coefficient of i related nodes; η represents the economic coupling coordination degree, and the control result under the control of the control coefficient fa [6]. The degree of coupling refers to the degree of influence between the two. Here, the degree of coupling between higher education and the regional economic system can be obtained by calculation. Let us denote it as η and η ∈ [0, 1].The larger the value, the higher the coupling level, and the smaller the value, the more uncoordinated. The coupling level is lowest when the η value is zero and highest when the value is one. Set the sequence parameter of the higher education subsystem as V1 and the sequence

180

S. Wang

parameter√ of the regional economic subsystem as V2 , the specific formula is as follows: η = 2V1V+1 ·VV22 . The degree of coupling can determine the degree of synergy between the two subsystems, but it is difficult to analyze the overall effect and comprehensive synergy of the two subsystems based on the degree of coupling. Therefore, a new variable T needs to be introduced, and its calculation formula is T = 0.5V1 + 0.5V2 . In this way, the coupling coordination degree function is constructed, thereby eliminating the disadvantages of the coupling function. Let √ B be the coupling coordination degree, and its calculation formula is as follows: B = η × T . For the coupling coordination degree, it can be divided into ten different grades in turn, as detailed in Table 2. Table 2. Evaluation criteria for coupled coordinated scheduling rating Coupling coordination B

0.00–0.09

0.10–0.19

0.20–0.29

0.30–0.39

0.40–0.49

Coupling level

Extreme disorders

Severe disorders

Moderate disorder

Mild disorder No disorders

Coupling coordination B

0.50–0.59

0.60–0.69

0.70–0.79

0.80–0.89

0.90–1.00

Coupling level

No coordination

Primary coordination

Intermediate level coordination

Good coordination

Senior coordination

After the construction of the index system, the collection and processing of the index data, the data for empirical analysis are obtained, including the order parameters of the higher education subsystem, the order parameters of the regional economic system, the coupling degree and the coupling coordination degree between the higher education and the regional economy, as shown in Table 3.

Design of Higher Education Aided Teaching System

181

Table 3. Data calculation results Years

Coupling degree A

Coupling coordination degree B

Coupling level

2005

0.960

0.203

Moderate disorder

2006

0.989

0.275

Moderate disorder

2007

0.999

0.288

Moderate disorder

2008

0.989

0.331

Mild disorder

2009

0.995

0.350

Mild disorder

2010

0.999

0.388

Mild disorder

2011

0.999

0.427

Endangered

2012

0.998

0.464

Endangered

2013

0.995

0.531

Barely coordinate

2014

0.999

0.548

Barely coordinate

2015

0.997

0.598

Barely coordinate

2016

0.992

0.630

Primary coordination

2017

0.992

0.657

Primary coordination

Set the index system of coupling system and coordinate the allocation of auxiliary teaching tasks, and the specific index system is shown in Table 4. Table 4. Index System of Coupled Systems Coupling Systems

Indicator level I

Regional Higher Education Development System

Higher education Talent Training in Colleges and Universities Scientific research in higher education Educational benefits

Regional economic development system

Economic aggregate Domestic and foreign trade volume Economic benefit level

Under the control of the load, the auxiliary system adjusts the operating parameters of the server according to regular or irregular time conditions, and issues corresponding query requests to the database control unit according to the choice of students and teachers, thereby achieving balanced distribution of queries The purpose of the task.

182

S. Wang

3.2 Setting up the Distributed Filtering Method for Teaching Content by Auxiliary System On the basis of achieving a balanced distribution of auxiliary teaching tasks, the distributed processing method of the auxiliary system is used to process massive amounts of higher education teaching data, and the original program is converted into numerous sub-processing programs in the same processing time to speed up the frequency of data processing. This technique presets the Euclidean distance between subroutines:

n  2 (4) Pμ r − Pμ s dμ = s=1

In the formula: dμ represents the Euclidean distance between any two adjacent subroutines whose load is μ ; Pμ r and Pμ s represent the load variables of program r, s and μ . Set the distance matrix at this time to W and determine the minimum distance element dmin . When the corresponding position of the data in the subroutine is i, j and its value is less than the threshold Z, then the distance matrix is merged to reduce the (W ) (W ) dimension [7]. In the setting program, the two nearest data class clusters are Ki , Kj ,   (W ) (W ) (W ) , so the new classification and the merged new class cluster is Kij = Ki , Kj (W )

(W )

(W )

is K1 , K2 , . . . , Km , and the system sets a distributed data processing subroutine according to the new class cluster. Based on the completion of the distributed processing program, the massive teaching data is classified linearly, so that the data in each processing unit is of the same type, or has the same purpose, or has similar value. Randomly select two sample data and divide them into two categories. The n-dimensional vector x represents the corresponding data features, and y represents the corresponding classification mark. The calculation expression of the linear classification hyperplane and classification function is:  T w x +e = 0 (5) f (x) = wT x + e In the formula: wT represents the transposed matrix established for the data type; e represents a fixed constant; f (x) represents the classification function. When f (x) > 0, the corresponding classification flag y = 1; when f (x) < 0, y = −1; when f (x) = 0, the support vector of the data is above the hyperplane. According to the classification of the above formula, it can be known that the linear classification is shown in Fig. 3 as follows. The triangles and circles in the graph represent vectors in two randomly selected samples [8]. Set constraints and establish a fast processing function to achieve instant processing of massive teaching data. The calculation expression of this function is as follows: βh λ=

n

vi f (x)

i=1

ϕ ln t(n−1)

(6)

Design of Higher Education Aided Teaching System

183

Fig. 3. Schematic diagram of linear classification

In the formula: λ represents the instantaneous processing function established; β represents the total amount of data; h represents the error correction coefficient; f (x) represents the processing constraint; vi represents the limit of the processing frequency in the i data segment; ϕ represents the control harmonic coefficient; t represents the instantaneous response time; n represents the processing path. According to the above formula, a mass data processing program is set to control the processing process of the subroutines, and distributed filtering processing of mass higher education data is realized. 3.3 Implementing Interactive Feedback in Teaching Systems The core of the designed higher education auxiliary teaching system lies in the design of interactive functions. The behavior interaction module is assigned to the computer control center, and the operation processing order of the interaction module is distinguished in the form of a flowchart. According to the preset teaching system framework, arrange the operating mechanism of the teaching system, and use the response function to complete the triggering of complex teaching procedures, data transfer, and layout conversion. The calculation expression for this response function is: f (w) = ks − λwi

(7)

In the formula: f (w) represents the response function under the w teaching program; k represents the conventional coefficient of the software response; s represents the level of the teaching materials inquired; wi indicates the comprehensive response coefficient of the w software teaching page under the selected i th operation target. This function is used to trigger the operation procedure when the user inquires the teaching content, so as to quickly enter the teaching content storage database according to a series of chain reactions. After the user enters the teaching database, by switching keywords, he sends instructions to the control center and obtains teaching information fed back from the system control center [9].

184

S. Wang

A new layered implicit algorithm is introduced, and the interaction data obtained is divided into three layers using this algorithm: weak feedback, strong feedback, and extremely weak feedback. It can be seen that there is U ∩ V , U ∩ W = ∅ between them, where W means very weak feedback; V means strong feedback; U means weak feedback. The comprehensive calculation formula of the implicit algorithm is:    h

ln q shi − shj × εij + f (w)δ  2   τ =− ωc Vij i=1,j=1

(8)

In the formula: q represents the total amount of specific information in the feedback layer; h represents the trust of the same type of information source; i, j represents any two layers of feedback; shi represents the amount of interaction behavior at the layer; shj represents the amount of interaction behavior data at the j layer; εij represents the differentiation coefficient between the two layers; δ represents the maximum likelihood parameter of f (w); represents the feedback decomposition estimated value. Use the feedback code to feedback the analysis results, some of which are shown in Fig. 4 below [10].

defupload_file_to_hdfs(filepath): client = InsecureClient(config.HDFS_URL) now = datetime.now() today_dir="%s/%02d/%02d" % (now.year, now.month, now.day) data_exist = client.status(today_dir, strict=False) if not path_exist: client.makedirs(today_dir) client.upload(today_dir, filepath) Fig. 4. Key feedback code

By relying on the above code, the obtained feedback data is uploaded to the feedback page through the data transfer unit, and the higher education assistant teaching system based on the economic coupling coordination analysis is designed.

4 Testing Experiment In order to verify the function of the designed higher education auxiliary system, a comparative experiment is proposed to compare the functional differences between the designed auxiliary system and the two traditional designs. This test work, according to the differences between the two, improve the original system of educational assistance function, the traditional design of the problems in the auxiliary system, make adjustments and modifications.

Design of Higher Education Aided Teaching System

185

4.1 Experiment Preparation Before the experiment starts, it is necessary to arrange the test environment in an orderly manner, according to the response time, system function, select the appropriate experimental object, so as to make a detailed experimental plan, and put forward more targeted contrast content. In the experimental test environment built this time, the server has 512 GB of memory, 2 TB of hard disk size and Windows Server 2018b of operating system; the server of the test machine contains 32 GB of running memory, 2 TB of hard disk, the operating system version is Windows 10, and the browser is IE11.0. Set the number of network nodes in this test experiment to 2000, and the node IP is the 16A class subnet mask in the same network; the data is stored in a decentralized form. Figure 5 below shows the basic environment for this experiment.

Fig. 5. Experimental test basic environment

The VRML browser is loaded into the computer of the selected three teaching systems. In order to ensure that the system can play the teaching content normally, download the Cycore series plug-ins to ensure that the dynamic teaching files can be opened normally. After the software of the three computers is installed, the teaching system is run and the experimental environment is tested. If the system can run normally after 1 h, there is no phenomenon of unstable operation and wrong screen casting, which indicates that the setting of the experimental environment meets the requirements of the experiment. The test was carried out in three groups at the same time, and in order to ensure the reliability of the output results, feedback analysis of the interactive results of the query search data of a certain subject in a certain grade is carried out, and the test results are shown in Fig. 6 below.

186

S. Wang

Fig. 6. Comparison of experimental results

Design of Higher Education Aided Teaching System

187

Analysis of the above three groups of experimental results shows that the educational aid system based on the economic coupling coordination analysis has feedback degree similar to the interactive result feedback for the 10 groups of teaching test data. Under the traditional design, the feedback results of the auxiliary teaching system (1) are high and low, and the feedback of the same set of interactive data will fluctuate violently, which shows that the feedback stability of the system is poor. The interactive data feedback curve has more stable data feedback in the first five groups of data, and the following five groups of data feedback results are extremely close to 0. According to the above analysis, the interactive feedback effect of the designed higher education assistant teaching system is better and the system is more stable.

5 Conclusion Based on the design of the traditional teaching system, the higher education auxiliary teaching system designed this time enhances the feedback of interactive results through the analysis of economic coupling coordination degree, and improves the stability of the auxiliary teaching system. However, the design of the higher education auxiliary education system proposed this time is carried out under idealized conditions. In the actual application process, the compatibility between other hardware equipment and system software must also be considered. In future research, Focus on research and optimization in this regard.

6 Fund Projects The 13th Five Year Plan of Jiangxi Educatian Science “Research on the Coordinated Development of Regional Higher Education and Economy” (17ZD065).

References 1. Devlin, M., McKay, J.: Teaching inclusively online in a massified university system. Widening Part. Lifelong Learn. 20(1), 146–166 (2018) 2. Pokhodnya, K., Anderson, K.J., Kilina, S.V., et al.: The mechanism of charged, neutral, monoand polyatomic donor ligand coordination to perchlorinated cyclohexasilane (Si 6 Cl 12). J. Phys. Chem. A 122(16), 4067 (2018) 3. Auerbach, J.E., Concordel, A., Kornatowski, P.M., et al.: Inquiry-based learning with RoboGen: an open-source software and hardware platform for robotics and artificial intelligence. IEEE Trans. Learn. Technol. 12(3), 356–369 (2018). PP(99): 1–1 4. Gao, M., Xu, G., Liu, T.Y., et al.: Development of teaching training and assessment system for warming acupuncture. Zhongguo Zhen Jiu = Chin. Acupunct. Moxibustion 39(9), 1021–1023 (2019) 5. Minati, L., Frasca, M., Yoshimura, N., et al.: Versatile locomotion control of a hexapod robot using a hierarchical network of nonlinear oscillator circuits. IEEE Access 6(99), 8042–8065 (2018) 6. Zhang, X., Wang, H.: Optimal dispatch method of transmission and distribution coordination for power systems with high proportion of renewable energy. Dianli Xitong Zidonghua/Autom. Electr. Power Syst. 43(3), 67–75 and 115 (2019)

188

S. Wang

7. Lai, C.M., Cheng, Y.H., Hsieh, M.-H., et al.: Development of a bidirectional DC/DC converter with dual-battery energy storage for hybrid electric vehicle system. IEEE Trans. Veh. Technol. 67(2), 1036–1052 (2018) 8. Wang, R., Xiao, F., Zhao, Z., et al.: Investigation on measurement method of transformer winding AC resistance by using auxiliary winding. Diangong Jishu Xuebao/Trans. Chin. Electrotech. Soc. 34(2), 245–254 (2019) 9. Li, M.: Example-based learning using heuristic orthogonal matching pursuit teaching mechanism with auxiliary coefficient representation for the problem of de-fencing and its affiliated applications. Appl. Intell. 48(9), 2884–2893 (2018) 10. Ratzinger, D., Amess, K., Greenman, A., et al.: The impact of digital start-up founders’ higher education on reaching equity investment milestones. J. Technol. Trans. 43(3), 1–19 (2018)

Research on the Optimization of the Allocation of Educational Resources in Modern History Based on Demand Features Hong-Gang Wang1(B) and Jing Xia2 1 College of Marxism, Fuyang Normal University, Fuyang 236037, China

[email protected] 2 The Research Institute of Informatization and Industrialization Integration,

China Academy of Information and Communications Technology, Beijing 200001, China

Abstract. In the past optimization methods of modern history education resource allocation, there is a problem of large financial cost consumption. Therefore, the optimization method of modern history education resource allocation based on demand characteristics is proposed. Based on the analysis of the actual needs of different users for educational resources of modern history and the characteristics of educational resources they need, this paper constructs the allocation model of educational resources of modern history, optimizes each link of the model in the process of construction, optimizes the operation mechanism of educational resources allocation at the same time, implements the allocation model of educational resources in the optimized operation mechanism, and achieves the reasonable allocation of educational resources The purpose of the source. The experimental results show that: compared with the traditional optimization method, the designed optimization method based on the demand characteristics of modern history education resource allocation consumes less financial cost, which is suitable for practical projects. Keywords: Demand characteristics · Modern history · Educational resources · Allocation optimization

1 Introduction The level of education development is a relatively abstract concept, which is affected by many factors such as economy, society, and history. It is an ideal method to determine the level of education development through direct measurement of education results. The indicators that reflect education results mainly include the student’s test results, the number of students at school, the enrollment rate, the graduate promotion rate, or the graduate employment rate, etc.]. However, in reality, due to the delayed, long-term, and potential of educational benefits, educational results are often difficult to accurately measure directly.

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 189–199, 2020. https://doi.org/10.1007/978-3-030-63955-6_17

190

H.-G. Wang and J. Xia

As the population is naturally distributed, there is no significant difference in the intelligence level of students in different regions, that is, there is no significant difference in the quality of students in different regions [2]. Therefore, although the richness of educational resources can not fully represent the level of educational development, it is undeniable that there is a strong correlation between educational results and educational resources. It can be said that the greater the quantity and quality of education resources in a region, the better the local education results [3]. Therefore, people usually indirectly reflect the level of education development through the measurement of educational resource indicators such as the investment in education funds, the allocation and use of teaching equipment, the composition and treatment of teachers, and the ratio of students to teachers. Educational resources, like all other resources, have a relationship between supply and demand [4]. The supply of educational resources is mainly provided by governments at all levels of the country, social organizations or individuals, and students and families to the sums of all levels within a certain period of time; the demand for educational resources refers to various types of educational institutions at all levels in order to conduct normal teaching activities and improve Institutional resource level and education quality, and the amount of resource input required for educational reform and innovation [5]. The continuous supply of educational resources is a prerequisite for the smooth development of educational activities. However, in reality, there is always an objective contradiction between the limited supply of educational resources and the unlimited demand for educational resources. To make basic education resources balanced in supply and demand, the following conditions must be met: First, social resources or education The input of total resources into basic education should meet the actual needs of the development and reform of basic education; the second is that the resource supply of schools in all regions should gradually stabilize at a scientific and reasonable level without too much fluctuation; the third is the supply of resources There are no shortages or excesses [6]. Therefore, it is necessary to study the optimization of the allocation of educational resources, especially for modern history. Nowadays, it is becoming more and more difficult to learn and popularize modern history. Studying the optimization of the allocation of educational resources for modern history for the study of modern history More helpful. In the traditional optimization method of modern history education resources allocation, there is too much waste of financial resources, which has a certain impact on the local economic development. Therefore, this paper puts forward the optimization method of modern history education resource allocation based on the demand characteristics, which solves the problems existing in the traditional optimization method.

2 Optimization of Educational Resources Allocation of Modern History Based on Demand Characteristics 2.1 Analysis on the Demand Characteristics of Modern History Education Resources In the optimization of educational resource allocation, the implementation of specific educational resource allocation optimization is often customized based on user needs.

Research on the Optimization of the Allocation

191

However, users cannot fully understand the constraint dependencies in the entire model, and can only choose independent features through a variety of other requirements and other information [7]. What features the user decides to include or remove is not relevant. Maybe the feature that the user specifies to be excluded is exactly what an included feature depends on. Therefore, analyzing and modeling the relationship between user needs and feature models is the key to solving the optimization of educational resource allocation based on demand features. Suppose the user decides that the feature of the final educational resource is Q, and the removable feature is temporarily empty. From the perspective of division, there are 16 feature combinations containing feature Q (24 ); from the perspective of configuration rules, all effective educational resources containing feature Q must include feature A, so adding feature A can reduce the feature combination from 16 to 16. 8. Among these eight feature combinations, there are only four valid feature combinations that satisfy the configuration rules, which are (q1 , q2 , a1 , a2 ), (q1 , q2 , a1 ), (q1 , a2 , a1 ), and (q1 , a1 ), respectively. From the above analysis, it can be seen that the specific needs of users often meet about 2n product variants (n is the number of features not specified by users). Further analysis shows that the common part of all effective configuration sets (such as feature Q and feature A) is the least dependent feature set including the user’s hard requirements; the features of non common parts can be bound according to the user’s actual needs or other constraints such as environment, such as feature q1 , q2 , a1 , a2 . Therefore, the minimum feature set including the user’s hard requirements is automatically calculated, and then its behavior is verified, which can lay the foundation for the subsequent dynamic binding based on the user’s requirements. Based on this, the common part feature set of the educational resource allocation set corresponding to the user’s needs is defined as the dependent feature set of the feature model slice. Assuming that the user decides that the final product contains feature set {q, a}, according to the above analysis, we can get that the dependency feature set is {q1 , q2 , a1 , a2 }. In addition, according to the decomposition relationship of features, features q1 and a1 are decomposition patterns satisfying XOR group, that is, they are mutually exclusive. However, the dependency feature set {q1 , q2 , a1 , a2 } only reflects the dependency between features, not the mutual exclusion between features q1 and a1 . On the other hand, users can specify which features are excluded by the final educational resources. To sum up, the optimization of educational resources allocation based on feature model is to realize the optimization of educational resources allocation in modern history through the combination and restriction of features. The user’s rigid demand reflects the three relationships of feature selection, abandonment and uncertainty. The dependent feature set only represents the public part of relevant educational resources, and the complete user needs need to be described by an exclusive feature set. Only the combination of dependent feature set and exclusive feature set can provide a more accurate analysis foundation for the subsequent realization of the abstraction of part of feature set in the whole behavior model. Therefore, the mapping result of user’s hard requirements to feature model is defined as slice result set, which consists of two parts of information: dependent feature set and exclusive feature set.

192

H.-G. Wang and J. Xia

2.2 Construction of Educational Resource Allocation Model Based on Demand Characteristics Based on the analysis of user needs in the previous section, a modern resource allocation model for educational history is constructed. The defined feature model slice is a unary operation on the feature model W . The input of this operation is called the slicing criterion, that is, the user’s choice of features. The slicing criterion consists of the mandatory feature set and the removed feature set specified by the user. The output is called the slice result set, and it consists of a dependent feature set and a rejection feature set. The following constraints exist on the educational resource allocation model: ηin = {∩x|x ∈ hsl }

(1)

    H |y ∈ hsl = ∩ y

(2)

hsl = {z|z ∈ W ∧ hsl ⊆ z}

(3)

ηexcel among them:

In the formula, hsl represents all the educational resources that meet the slicing criteria, H represents a limited feature set in the allocation of educational resources, ηin represents a dependent feature set, and ηexcel represents an exclusive feature set. x, y and z represent regular variables. The above constraints are the results derived from the semantic perspective of the feature model, and the slice result set sl (W ) = ηin , ηexcel  is obtained. The result of feature slicing reflects the feature set which is most closely related to the specific needs of users. Based on this, we can further optimize the allocation of educational resources. The calculation of the feature slice result set needs to meet the needs of the user and the overall constraint information of the feature model. An algorithm designed is the slice result set algorithm. The input of the algorithm is the user’s demand for features. Set Nin and excluded feature set Nsl . The output is dependent feature set Pin and excluded feature set Psl . First traverse each feature in Nin , the specific operation is as follows: if the feature does not belong to the feature dependency set Pin and the feature exclusion set Psl , then it is merged into the set Pin ; at the same time, starting from the feature node in two directions The feature related nodes are searched. One direction is to its parent node until the root node is searched; the other is to its child node until the terminal node is searched. Any feature node in both directions will be included in the feature dependency set as long as it satisfies the dependency. In the same way, the algorithm traverses each feature in Nsl , and the traversal process is still performed in two directions. The difference is that the latter only searches its parent node when searching in the direction of its parent node, and only includes its parent node feature when the feature node satisfies the required attribute in the and group decomposition relationship; the search in the direction of its child node is the same as the feature dependency set until the terminal node is searched.

Research on the Optimization of the Allocation

193

According to the above search process, various types of educational resource allocation have been passed in modern history, and traversal is a characteristic of demand. This is the basis for educational resource allocation and the combination of operating mechanisms to achieve the purpose of optimization. 2.3 Optimize the Operating Mechanism of Education Resource Allocation The optimal allocation of modern history education resources is a dynamic process involving many factors such as objectives, environment, motivation and operation mechanism. These factors interact with each other and jointly affect the flow direction and final allocation effect of modern history education resources allocation [9]. From the perspective of system theory, this paper studies the mechanism of the optimal allocation of modern history education resources, analyzes the interdependence between the modern history education system and the environment, looks for the power source to promote the continuous optimal allocation of modern history education resources, and analyzes the operation mechanism and implementation path of the optimal allocation of modern history education resources [10]. The optimal allocation of educational resources plays an important role in the formation of the dissipative structure of the modern history education system and the realization of the balanced development of basic education. From the perspective of system science, that is, from closed to open, from chaos to order, from simple to complex. As a subsystem of the regional social system, the modern history education system has the basic characteristics of a general system, that is, it is composed of certain elements, has a certain structure and function, and exists independently of the surrounding environment. At the same time, it has an open character. Exchange material, energy and information with the environment. Specifically, the modern history education system refers to a network system composed of organizations, institutions, and implementation conditions related to the process of allocating resources for modern history education. It is composed of a number of regional governments and education departments, responsible for basic education information consultation and evaluation. It is composed of a supervision agency and several schools. The education system interacts with the surrounding environment. Under the promotion and pulling of various forces, various behavioral organizations in the system actively respond to changes in the environment, continuously formulate and adjust the development strategy of modern history education, and coordinate the relationship between basic education and regional environment, To promote the optimization of the allocation of educational resources in modern history. On the basis of the above contents, different mechanisms are adopted to optimize the allocation of modern history education resources. First, decision-making and consulting mechanisms. Decision-making is the premise and basis for scientific management of modern history education, and it is a key measure to ensure that local education is coordinated with regional economic and social development. Decision-making plays a central role in optimizing the allocation of educational resources in modern history. Effective decision-making can promote a reasonable and orderly flow between internal and external environmental elements, and ensure that human, financial, material and information resources of a certain size and quality are input into the modern history education system. Therefore, establishing a scientific and

194

H.-G. Wang and J. Xia

efficient decision-making mechanism for the modern history education system is of great significance for optimizing the allocation of modern history education resources. The rationality of organizational design is directly related to the scientificalness of decision-making. We must start with improving the organizational structure of decisionmaking, improve the education management system with strong overall planning and clear rights and responsibilities, and implement the hierarchical decision-making management of the central and local governments for basic education. In order to avoid the improper coordination and impassability of political orders caused by the differences of interests among governments at all levels, when designing the organizational structure and system of government decision-making, we must We should establish a clear and reasonable division of labor responsibility system to define their respective responsibilities, powers and interests, so as to make the behavior of the macro decision-makers coordinated. It is necessary to establish and perfect the micro decision-making body, make the school and the society’s strength fully play in the process of promoting the allocation of regional basic education resources, and provide a good system environment for all levels of education administrative departments and schools to make rational and independent decisions. Focusing on the transformation of government functions and the streamlining of administration and decentralization, we will improve the level of public education services, clarify the responsibilities of governments at all levels, unify the leadership and management of national education by the central government, formulate development plans, policies and basic standards, coordinate the management of basic education within the region by the provincial government, promote the balanced development of basic education, and implement the financial responsibility for the development of basic education in accordance with the law. Consultation is an auxiliary process of decision-making. Governments at all levels must set up education advisory committees to provide consultation evidence for education reform and development, and to improve the scientific nature of major education decisions. The second is the input and implementation mechanism. Education investment is a basic and strategic investment for the long-term development of the country, and an important function of public finance. We should improve the system of raising educational funds through various channels, with government investment as the main source. Governments at all levels should optimize the structure of financial expenditure and give priority to education as a key area of financial expenditure. In strict accordance with the provisions of education laws and regulations, ensure the “three growth” of education funds, and ensure the stable source and growth of school running funds. Compulsory education is fully included in the scope of financial security, and an investment system in which the State Council and the local people’s governments at all levels share the burden according to their duties and the provincial government is responsible for the overall implementation is implemented. The investment mechanism of government investment, social sponsor investment and family reasonable burden should be established in preschool education. Ordinary high school implements the mechanism of financial investment as the main and financing from other channels as the auxiliary. We will further increase investment in education in rural areas, remote and poverty-stricken areas and ethnic minority areas, and the central government will support the development of

Research on the Optimization of the Allocation

195

education in underdeveloped rural areas and ethnic minority areas by increasing transfer payments. Implementation is the process of implementing the plans and programs formed by decision-making. Whether it is the macro-decisions of governments at all levels or the micro-decisions of schools, education administration departments and schools must complete education activities. Therefore, the main body of implementation is the education administrative department and school. Only when the education administrative department and school earnestly complete and implement various decisions, and continuously innovate in the process of implementation, improve the efficiency of the allocation of educational resources, can promote the overall basic education system in the province Optimized allocation of resources. The third is the supervision and feedback mechanism. Supervision is the basic guarantee of efficient implementation of various decisions and the control means of implementation. Different implementation projects have different decision-making plans and evaluation basis, which requires the government, society and schools to take necessary management methods to evaluate and supervise the process of each component of the implementation link, so as to ensure that education activities are carried out within a certain range of standards. Governments at all levels should conscientiously perform the responsibilities of overall planning, policy guidance, supervision and management, and the provision of public education services, establish and improve the public education service system, gradually realize the equalization of basic public education services, and maintain education fairness and order. Establish and improve the basic standards of national education, integrate the national education quality monitoring and evaluation institutions, improve the monitoring and evaluation system, and regularly issue monitoring and evaluation reports. The coordination of the function of the education system and the output of the system is largely achieved through feedback, which is a controller for the optimal allocation of resources in provincial basic education and an important means for achieving sustainable development in modern history education. Adhere to financial management in accordance with the law, strictly implement the national financial fund management legal system, strengthen the supervision of the use of funds, strengthen the audit of the entire process of major project construction and use of funds, and ensure that the use of funds is standardized, safe and effective. Establish a performance evaluation system for the use of funds, strengthen the evaluation of the use of funds for major projects, establish and improve the management system for the allocation, use, and disposal of state-owned assets in schools, and improve the use efficiency. Through supervision and feedback, the goal of optimizing the allocation of educational resources in modern history is finally achieved. The fourth is the policy incentive mechanism. Incentive mechanism, also known as incentive system, reflects the interaction between incentive subject and incentive object through a rational system. The connotation of incentive mechanism is the elements of several aspects of the system. Once the incentive mechanism is formed, it will play an internal role in the organization system itself, make the organization function in a certain state, and further affect the survival and development of the organization. We should take promoting the scientific development of education as an important part of

196

H.-G. Wang and J. Xia

the performance appraisal of Party committees and governments at all levels, strengthen the legal construction of basic education, further improve the institutional framework of investment, management, use, audit and supervision of educational resources, gradually establish the system of announcement and examination of educational investment, and fundamentally guarantee the standardization of the allocation of basic educational resources. Do a good job of performance evaluation of basic education resource allocation, strengthen fund management, establish education fund supervision and management institutions, improve education fund supervision functions, establish education fund implementation analysis report system, establish fund performance evaluation system, and strengthen major project fund use appraisal To ensure the effective use of educational resources. Take the implementation of education investment and the use of funds as an important basis for the performance evaluation of governments at all levels and accept social supervision. Necessary rewards are given for meeting educational standards and high use efficiency. In summary, there is an organic relationship between the environment for the optimal allocation of education resources in modern history and the input mechanism, supervision mechanism, and incentive mechanism. The interaction between them is shown in Fig. 1, thereby realizing education in modern history. The purpose of resource allocation optimization.

economic

Geogra phy Material resources

Manpower

politi cal

financial resources population

system

decision making

advisor y Feedback

Supervi se

carried out

Incentives Educational resource allocation goals

Fig. 1. Operation mechanism of educational resource allocation optimization

Research on the Optimization of the Allocation

197

The educational resource allocation model optimized based on demand characteristics will be implemented under the above-mentioned operating mechanism to realize the optimization of educational resources allocation in modern history.

3 An Experimental Study on the Optimization Method of Educational Resources Allocation in Modern History 3.1 Experimental Platform Construction A lot of teaching resources need to be used in the experimental research of the optimization method of modern history education resource allocation, so an experimental platform based on Hadoop is built. The building process of Hadoop distributed environment is to simulate linux environment through cygwin64 under Windows operating system. Install jdk64 and Tomcat server, and prepare to start building Hadoop platform. On the server side, set the user name as administrator, one master node and two slave nodes; the deployment work is carried out on the master branch, slave1 and slave as nodes obey the configuration. Modifying the hosts on the master is mainly to change the calculation and system path to /etc. /hosts file, add the IP address of the master and slave nodes at the end of the hosts file, install and start the SSH service, and configure SSH password free login function, modify the configuration file of Hadoop, add the installation path of JDK in hadoop-env.xml file, and enter the start-all.sh command to start Hadoop. The experimental platform is completed. 3.2 Experimental Data Preparation In the experimental study of the optimization of the allocation of educational resources in modern history, the experimental data of the ratio of teacher-student resources of ordinary primary schools, junior high schools, and ordinary high schools in a certain province is used as experimental data (Table 1). Table 1. Teacher-student ratio data of basic education resources in a province Number

Primary school

Primary school

High school

Total

1

14.2:1

16.3:1

15.2:1

15.23:1

2

14.6:1

16.7:1

15.6:1

15.63:1

3

14.7:1

15.8:1

17.2:1

15.90:1

4

14.9:1

16.3:1

18.5:1

16.56:1

5

14.3:1

16.1:1

16.9:1

15.76:1

Based on the data of teacher-student ratio of basic education resources in the table, this paper estimates the financial cost of the allocation of education resources, and then

198

H.-G. Wang and J. Xia

optimizes it by using different optimization methods of the allocation of modern education resources, and analyzes the advantages and disadvantages of different optimization methods of the allocation of modern education resources according to the financial cost of the optimized allocation of education resources. 3.3 Experimental Results and Analysis The results of financial cost required after using different optimization methods of education resource allocation are shown in the following figure.

Fig. 2. Experimental results of different optimization methods

Observing the results in the figure, Fig. 2 (a) shows the results of the 5 groups. There are two groups of results that have reduced costs after optimization, and the other three groups have increased costs after optimization. Compared with the previous one, the cost after optimization has decreased significantly, which shows that the optimization

Research on the Optimization of the Allocation

199

method can effectively reduce the cost consumption. To sum up, the design method of resource allocation optimization based on the characteristics of modern history education consumes less cost, and this method is better than the traditional optimization method.

4 Concluding Remarks The allocation of educational resources in modern history is closely related to the balanced development of education, which affects the sustainable development of the overall level of education and socio-economic development. It is of great significance to study the optimization of the allocation of educational resources for the development of education and economic levels. Designing a method for optimizing the allocation of educational resources in modern history based on the characteristics of demand, solving the problem of large consumption of financial resources in traditional optimization methods, rationally allocating teaching resources, and promoting the long-term development of the education industry, while ensuring that the social economy is developing in a good direction. Fund Project. The Fuyang Normal University’s 2019 undergraduate teaching engineering teaching research project “MOOC + flip classroom mixed teaching model design”, project number 2019JYXM35. The Provincial University Student Innovation and Entrepreneurship Project of 2019 “Research on the Innovation and Development of Red Tourism in Anhui Province Based on SWOT Analysis”, Project No. S201910371062.

References 1. Guoqing, Z., Junhua, G., Ling, H.: Research on open university students’learning engagement based on NSSE. Educ. Res. 25(04), 91–99 (2019) 2. Wenxing, L., Chu, L.: Forecasting of short-time tourist flow based on improved PSO algorithm optimized LSSVM model. Comput. Eng. Appl. 55(18), 247–255 (2019) 3. Feng, M., Luyang, H., Xiaoling, H., et al.: Research on current situation of human resource allocation in MRI departments of H province. Chinese Hosp. Manage. 39(06), 25–27 (2019) 4. Huining, X., Hua, Q., Jianwei, : Characteristics of drivers’ information demand in autonomous vehicles. Urban Transport of China 17(03), 96–104 (2019) 5. Lifeng, H., Xiaomu, Z., Jia, L.: Exploring innovative information literacy instructions in higher education in the new environment: a case study of Tsinghua university library. Library and Information Service 62(24), 12–17 (2018) 6. Jian, W., Yan, T.: The evolution, characteristics and enlightenment of the pre-service PE teacher education system in china. J. Chengdu Sport Univ. 45(04), 98–104 (2019) 7. Fuhua, L.: Research on the endogenous demand of national strength in the construction of teachers’team. Tsinghua J. Educ. 39(06), 88–95 (2018) 8. Lihong, D., Libo, S.: On value, problems and strategies of teaching cost accounting in technology-applied undergraduate colleges. Vocat. Tech. Educ. 40(11), 16–19 (2019) 9. Weixuan, S., Tangqi, T., Shanggang, Y., et al.: The differentiation and effects of social-spatial accessibility in compulsory education resources in Nanjing. Geog. Res. 38(08), 2008–2026 (2019) 10. Feng, L., Yong, Q.: Research on current situation of jiangsu science and technology resource allocation based on clustering and corresponding analysis. Science & Technology Progress and Policy 36(10), 41–48 (2019)

Design of Learning Feedback System of Sports Training Based on Big Data Analysis Lei Chen1 , Fei Gao2(B) , and Gewei Zhuang3 1 Education College of Tibet University, Lhasa Tibet 850000, China

[email protected] 2 School of Information Science and Technology, Tibet University, Lhasa Tibet 850000, China 3 Power Science Research Institute of Shanghai Electric Power Company, Shanghai 200000,

China

Abstract. Due to the poor stability of the traditional sports training learning feedback system, it can not improve the efficiency of students’ sports training learning, thus reducing the learning interest of students. Therefore, a sports training learning feedback system based on big data analysis is proposed and designed. Through the positive feedback mechanism and the negative feedback mechanism, the study feedback mechanism of sports training is analyzed. Under the big data analysis, the study feedback mechanism of sports training is used. According to the function structure of the system, the study feedback system of sports training is designed, including the system login module, the data management module and the study feedback module. The experimental results show that the application of the designed learning feedback system based on big data analysis improves the efficiency of the conventional teaching by 34%, the learning interest of students by 39%, the satisfaction of students’ learning feedback, and the stability and retrieval rate of the system. Keywords: Big data analysis · Sports training · Learning feedback system · Feedback mechanism

1 Introduction With the rapid development of IT technology, all walks of life are facing the pressure of massive data processing. Human intelligence alone can no longer meet the needs of massive information computing and analysis. In 2012, the Obama administration released the big data research and development initiative on the White House website, aiming to improve the ability to acquire knowledge and insights using a large number of complex data sets. Six federal government agencies agreed to invest more than $200 million in this. With the rapid development of IT technology, all walks of life are facing the pressure of massive data processing. Human intelligence alone can no longer meet the needs of massive information computing and analysis [1]. In sports training, by using the learning feedback mechanism of sports training, students can obtain certain learning © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 200–214, 2020. https://doi.org/10.1007/978-3-030-63955-6_18

Design of Learning Feedback System of Sports

201

effects through their own efforts, usually in the form of teacher evaluation, examination results, student evaluation, etc. for such learning effects, they will return to the consciousness of learners, gradually form new information to regulate learning process, thus constantly learning feedback. In the learning process, a clear understanding of their own learning results will adjust the learning process of learners in a timely manner, constantly improve their enthusiasm for learning, enable them to get good grades, and promote students to increase their efforts. On this basis, students can understand their own shortcomings, so as to correct them in time. It can not only stimulate students’ interest in learning, but also strengthen learning motivation, which is the feedback effect of learning results. In the process of learning, middle school students master their own achievements, get some spiritual satisfaction, produce a kind of self-confidence, accompanied by a sense of pleasure, excitement and relaxation; at the same time, they further arouse interest in learning, show great enthusiasm for learning, and improve students’ enthusiasm and initiative in learning [2]. At present, domestic and foreign research in this area has made some achievements. Literature [3] designed an Internet-based Interactive Platform for physical education theory course information, which uses Windows Based on the web solution provided by IIS in XP, we collect all kinds of information needed by the system, use Dreamweaver, flash, ASP and other dynamic Internet development tools, integrate all kinds of physical education theory course information systems according to the needs, take the dynamic hypertext format as the basis of information platform, and provide the information of physical education theory course through the way of network information interaction Users. The research shows that the data of the platform has the characteristics of integration, control and interaction. The platform has the functions of knowledge diagnosis, online interaction, dynamic management, auxiliary teaching and network distance teaching. However, the stability of the system is poor, which can not improve the efficiency of sports training and learning, so as to reduce students’ interest in learning. Literature [4] designed a general video feedback system suitable for college physical education teaching. Based on the analysis of the challenge of college physical education teaching, based on the design theory of the feedback video system and the analysis of the application examples of the system, according to the different needs of different users, the complete functional framework of the general feedback video system is given. However, students’ satisfaction with the system is low, and the retrieval rate of learning feedback data is reduced. To solve these problems, this paper designs a learning feedback system of sports training based on big data analysis. Big data analysis is a widely used Internet technology at present. Applying it to the feedback design research of sports training learning is conducive to the improvement of the whole sports training teaching quality.

2 Research on Big Data Analysis Big data has the characteristics of large amount of data, complex data structure, fast data generation and low data value density, which increase the difficulty of effective analysis of big data. Big data analysis has become the core content of current exploration of big data development. Therefore, it is necessary to conduct in-depth analysis of the connotation and extension of big data analysis. Big data analysis is the product

202

L. Chen et al.

of rethinking data science and exploring new models in data intensive environment. Strictly speaking, big data is more like a strategy than a technology. Its core idea is to manage and extract value from massive data in a much more effective way than before. Big data analysis is the core of big data concept and method. It refers to the process of analyzing massive data (i.e. big data) with various types, rapid growth, and real content, and discovering useful information such as hidden patterns and unknown correlations that are helpful for decision-making. Therefore, this paper believes that big data analysis is a data analysis process that widely collects and stores data according to the data generation mechanism, formats and cleans the data, based on the big data analysis model and supported by the integrated big data analysis platform, uses the cloud computing technology to schedule the computing and analysis resources, and finally excavates the patterns or laws behind the big data [5]. The big data analysis process is shown in Fig. 1. Start Cloud storage model

Cloud service interface Y

N Safety certificati on

Cloud data center

Programming model

Parallel database

Network access

End

Fig. 1. Big data analysis process

3 Analysis on the Feedback Mechanism of Sports Training Learning The feedback mechanism of sports training based on big data analysis is based on cybernetics, information theory, systematology and modern teaching theory. Through information feedback as the main channel, it penetrates the classroom with active and positive learning attitude and advocates multi-directional modern teaching information transmission. It changes the single way of information transmission in traditional teaching and “imitation one” in traditional teaching Memory is a simple and inefficient learning habit. Students follow the learning activity law of “exploration, memory and creation”; with

Design of Learning Feedback System of Sports

203

information feedback as the main channel, in the teaching, on the basis of changing the traditional teaching, teachers master the basic knowledge, basic technology and basic skills, take teachers as the leading role and students as the main body, realize teachers’ scientific and reasonable control skills of teaching activities, and achieve the teaching purpose of cultivating students’ comprehensive quality. In order to give full play to the overall function of the teaching system and realize the purpose of teaching and educating people, the feedback mechanism of sports training and learning always takes information feedback as the main line, regulates teaching and learning in time and strives to obtain the best teaching effect. The so-called system is an organic whole composed of several interdependent and interdependent elements for the realization of a certain function. According to the principle of system theory, teaching, as a practical activity, is a relatively independent system, which has many interrelated elements such as teachers, students, teaching materials, teaching methods, learning methods, etc. These elements are interdependent in teaching activities. Teachers can’t do without students, students can’t do without teachers, teachers and students can’t do without textbooks, without teaching methods, teachers can’t follow the rules, without learning methods, students can do twice as much as they can. At the same time, these elements are mutually restricted. Teachers can’t teach at will, and they must do according to students’ needs. Students’ learning methods can’t do at will, and they must accept teachers’ management and guidance. Teachers’ teaching methods should be based on students’ actual aptitude, and students’ learning methods should be selected according to their own characteristics. In teaching activities, these elements must cooperate with each other to achieve their overall functions and improve the quality of teaching. However, in traditional teaching, it is not conducive to give full play to the overall function of the teaching system to attach importance to the leading role of teachers, ignore the main position of students, attach importance to teaching methods and ignore learning methods. In view of these disadvantages, the feedback mechanism of sports training and learning fully mobilizes the enthusiasm of students’ active learning, guides the learning methods, and makes students’ development lively. The learning feedback mechanism of sports training includes positive feedback mechanism and negative feedback mechanism. When students get good academic performance or get good evaluation, their enthusiasm will improve, which is a positive feedback mechanism; when students’ academic performance is not good or get bad evaluation, their enthusiasm will be low, which is a negative feedback mechanism. Many students lose confidence in learning because of their poor performance. They get feedback every time, but they don’t know how to adjust their learning state. Psychologists Rosie and Henri further showed that the feedback mechanism is particularly effective in learning, especially the daily feedback is more efficient than the weekly feedback. The learning feedback mechanism of sports training emphasizes the timely feedback and regulation of information, which is the superiority of this mechanism. The feedback is divided into three ways: pre feedback, immediate feedback and delayed feedback.

204

L. Chen et al.

Pre feedback refers to before the implementation of teaching design. According to the feedback information of the previous teaching process, the teaching designer analyzes the teaching background, the existing level of students’ knowledge and ability, the psychological preparation of students’ learning new knowledge and other factors, so as to formulate the goals and Strategies of teaching design. Immediate feedback is to give feedback immediately when students learn new knowledge and have some behavior changes. Because the students have fresh memories of the reaction they just made, they can strengthen the correct response and correct the wrong response in time through immediate feedback, so as to achieve the purpose of enhancing the teaching effect. Delayed feedback is to give feedback after students have made a learning response. Experienced teachers often take a “cold treatment” approach to problems that are difficult to dredge students’ ideological understanding for a while, which is delayed feedback. This way of dealing with students left time for thinking, more conducive to the solution of the problem. In the process of feedback teaching, teachers, students and teachers and students are constantly exchanging and feedback information. A closed feedback loop is formed between teachers and students, and the two sides influence each other. Since sports training is a bilateral activity between teachers and students, teachers should have strong professional ability to output enough information. Whether the information input by teachers can achieve the expected effect depends on the ability of students to accept the information. The difference of learning effect is accompanied by the difference of receptive ability and feedback in different ways. For information feedback from different aspects, the teacher should make use of, analyze and revise it in time, and then pass the new information to the students after completing this process. It can make the feedback information input, output and control continuously, thus forming a closed whole sports training learning feedback system [6]. Physical education teachers deepen the understanding of feedback teaching in the process of teaching. According to different learning stages, different learning bases and different course content departments, timely and accurately obtain information, carry out comparative analysis on the preparation content, and adjust and change the teaching methods accordingly. For example, when students have psychological fear difficulties in teaching, we should consider reducing the difficulty of teaching and so on. At the same time, it has significant effect to strengthen the relationship between teachers and students in teaching, to adjust the internal, to obtain positive information, and to promote feedback to develop in a positive direction. In the teaching process, the effective information exchange and timely feedback between teachers and students can not only enhance the cooperation between teachers and students, realize the mutual improvement of teaching and learning, but also effectively improve the teaching effect [7].

Design of Learning Feedback System of Sports

205

4 Learning Feedback Design of Sports Training Based on Big Data Analysis 4.1 Functional Structure of the System Based on the above analysis of the learning feedback mechanism of sports training, this paper designs a learning feedback system of sports training based on big data analysis. The setting of the user’s authority of the sports training learning feedback system ensures the security of the system; the database query function enables the user to modify the data in the database directly without opening the database, which is the biggest advantage of the system. The sports training learning feedback system also has the extremely convenient database backup and recovery functions. The functional structure of the sports training learning feedback system under big data analysis is shown in Fig. 2:

Learning feedback module

Learning feedback system of sports training

System login module

Data management module

Fig. 2. System function structure diagram

4.2 System Login Module In order to increase the security of data, the system adds a login module. Before using the system, the user must have the login permission, otherwise the system will not run, which can prevent the personnel without access to the database from modifying the data of the database. The system login interface is shown in Fig. 3.

206

L. Chen et al.

Fig. 3. System login interface

4.3 Data Management Module The main function of this module is to display the data, which is stored in the Microsoft SQL Server 2000 database, in the form. If necessary, you can modify the data in the form grid directly to modify the data in the database. Because of the implementation of synchronous operation, what users see is the data in the database. In addition, the system also provides modification function - through this interface, users can add data, modify data and delete data to the database. 4.4 Learning Feedback Module Using feedback teaching to stimulate the desire for knowledge shows a positive learning motivation. Because of continuous learning feedback, information flow interacts between learners and environment, forming a loop. As shown in Table 1: Table 1. Learning feedback of sports training based on big data analysis Influence factor

Important advice Other

Experiences influence

91.7%

The influence of the class 87.9%

8.3% 12.1%

School attention

83%

17%

Teachers influence

85.9%

12.3%

Sports habit

82.8%

17.2%

The first level of feedback is mainly from classroom teaching, which is the most basic and timely information feedback. The purpose is to make teachers know more about the overall situation of students, master their physical, psychological quality, technical skills and personality differences, and improve teaching methods in time to improve the classroom teaching effect [8].

Design of Learning Feedback System of Sports

207

The second level feedback comes from the semester assessment. Through the assessment and inspection, it summarizes the problems in teaching in this stage, adjusts the teaching schedule and plan in time, improves the teaching method, and puts forward a reasonable and scientific teaching control plan [9]. The third level feedback mainly summarizes the effect of the implementation of the syllabus in the whole sports training through the completion examination of physical education class, finds out its shortcomings, so as to adjust the requirements and proportion of the syllabus, and puts forward more perfect contents of the syllabus, so as to implement the optimized teaching control process [10]. The fourth level of feedback mainly comes from the information feedback of students’ sports after graduation. It can feed back some knowledge that teachers are eager for in university sports training. To some extent, it guides the formulation and modification of teaching objectives and syllabus, provides reliable basis in teaching key points, and points out the direction for how to cultivate students’ sports interest and form lifelong sports concept in university stage. The learning feedback module is shown in Fig. 4:

Data analysis results

The database

Model library

The knowledge base

Fig. 4. Learning feedback module

According to the designed learning feedback system of sports training based on big data analysis. Design the learning feedback process of sports training, as shown in Fig. 5:

208

L. Chen et al.

Start Control system Knowledge required

Capacity requirements

N

Y Problem design N

Production monitoring N evaluation Y

Feedback system Y

Feed management The result of teaching

End Fig. 5. Design process of learning feedback in sports training

5 Example Analysis In order to ensure the effectiveness of the learning feedback system based on big data analysis, it is necessary to analyze its effectiveness. In the test process, different classification detection methods are taken as the test object, and the optimization ability of detection methods is analyzed, and simulation analysis is carried out with different specifications of test objects. In order to ensure the accuracy of the experimental process, we need to set the parameters of the experiment. In this paper, the simulation test uses the test data as the test object, uses three different design methods to carry out the classified test, and analyzes the simulation test results. Because the analysis results and analysis methods obtained by different methods are different, it is necessary to ensure the consistency of test environment parameters in the test process. The test data setting results in this paper are shown in Table 2:

Design of Learning Feedback System of Sports

209

Table 2. Comparison of students’ physical fitness Group

The experimental group The control group

Leapfrog

9.7

9.3

Reverse layup

9.6

9.4

Music leg go

9.8

9.5

Walking

16.8

15.8

Running

17.5

16.4

According to the above experimental parameters, this system is used to compare the satisfaction of students’ learning feedback with literature [3] system and literature [4] system, and the comparison results are shown in Tables 3, 4, and 5. Table 3. A systematic comparison of students’ satisfaction with learning feedback in this paper Survey questions

Satisfaction%

Curriculum Provision

93.6

The way for school

96.4

Evaluation way

97.5

Classroom interaction

91.2

The teaching evaluation 92.8

Table 4. A comparison of the satisfaction of students’ learning feedback in literature [3] Survey questions

Satisfaction%

Curriculum Provision

82.6

The way for school

80.3

Evaluation way

84.5

Classroom interaction

86.2

The teaching evaluation 82.6

It can be seen from the table that there is a significant difference in the examination results of the experimental class and the control class after the experiment. The reason is that there are many factors affecting the sports results, such as physical fitness, the length of contact time, the exertion of technical level and so on. For college students,

210

L. Chen et al.

Table 5. A comparison of the satisfaction of students’ learning feedback in literature [4] system Survey questions

Satisfaction%

Curriculum Provision

77.2

The way for school

76.5

Evaluation way

72.6

Classroom interaction

74.8

The teaching evaluation 79.5

through the test before the experiment, we can see that there is not much difference in physical quality. The main reason that affects their sports performance is the sports time, sports technology and skill proficiency. And the satisfaction of students’ learning feedback using this system is higher than that of literature [3] system and literature [4] system. The application of learning feedback mechanism in sports training teaching in Colleges and universities can stimulate students’ interest in learning, arouse students’ enthusiasm in learning, and help students to increase contact time in and out of class and improve their sports performance. Through the timely feedback of information, teachers can check the completion of teaching objectives, while students can also self-examine through the feedback information in class and the feedback evaluation sheet filled in after class to find out the deficiencies so as to increase the practice time and times and improve the sports performance. In order to further verify the effectiveness of this system, we use this system, literature [3] system and literature [4] system to compare and analyze the retrieval rate of students’ learning feedback data. The comparison results are shown in Fig. 6. According to Fig. 6, the retrieval rate of student learning feedback data in this system is up to 75%, while that in document [3] system and document [4] system is up to 52% and 49%, respectively. The retrieval rate of students’ learning feedback data in this system is higher than that of literature method.

Design of Learning Feedback System of Sports

Fig. 6. Comparison of students’ satisfaction with learning feedback in three systems

211

212

L. Chen et al.

In order to further verify the performance of this system, the stability of this system is compared with that of [3] and [4], and the comparison results are shown in Fig. 7.

Fig. 7. Stability comparison of three systems

Design of Learning Feedback System of Sports

213

Fig. 7. (continued)

According to Fig. 7, the stability of the system in this paper is up to 88%, which is higher than that of literature [3] and literature [4].

6 Concluding Remarks Based on big data analysis, this paper designs a study on the feedback design of sports training learning, which is conducive to improving the level of students’ sports technology, and can effectively improve students’ ability of creative thinking and observation, analysis and problem-solving. More conducive to creating a harmonious teaching environment, students help each other, exchange, and improve the teaching effect. It is conducive to the mastery of students’ theoretical knowledge and the improvement of sports skills, to fully reflect the main role of students in the classroom and the leading role of teachers, and to improve and develop students’ creative thinking, optimize teaching methods, enhance learning confidence, enrich evaluation methods, and strengthen the teaching effect.

References 1. Cassidy, T., Stanley, S., Bartlett, R.: Reflecting on video feedback as a tool for learning skilled movement. Int. J. Sports Sci. Coaching 12(3), 513–520 (2017) 2. Aiken, C.A., Fairbrother, J.T., Post, P.G.: The effects of self-controlled video feedback on the learning of the basketball set shot. Adv. Mater. Res. 29, 947–957 (2017) 3. Wang, Y., Ji, Y., Dong, L.: The information interaction platform development of sports teaching theory course based on internet. Third Int. Conf. Educ. Technol. Training (ETT) 3, 210–219 (2017)

214

L. Chen et al.

4. Zhang, B.F., Xiong, Y.: The design of general video feedback system suitable for college sports teaching. Adv. Mater. Res. 71, 786–790 (2017) 5. Tesser, A.: The important of heritability in psychological research. Pshchological Reviw. 82(6), 135–141 (2017) 6. Yao, L., Wen, Y.: Theoretical thinking on evaluation of physical education teaching in China. J. Beijing Sport Univ. 33(10), 163–172 (2017) 7. Deguang, Z.: Discussion on formative evaluation of physical education. J. Plain Univ. 12(6), 401–410 (2017) 8. Baohong, H., Dengyun, W.: Statistical analysis of index error in physical education evaluation. J. Wuhan Sports Univ. 36(12), 230–241 (2017) 9. Xiaoyu, X.: Construction of a new evaluation system for college physical education teaching. J. Adult Educ. College, Hubei Univ. 2(13), 12–26 (2017) 10. Jinhua, L.: Experimental study on improving the teaching effect of PE. J. Binzhou Teachers College 14(7), 304–310 (2017)

Design of Electrical Remote Control Teaching System Based on Intelligent Ubiquitous Learning Network Model Zhuang Gewei(B) Power Science Research Institute of Shanghai Electric Power Company, Shanghai 200000, China [email protected]

Abstract. The traditional teaching system of electric remote control remote control teaching problem of low precision and intelligence in ubiquitous learning network model is put forward for this electric remote control teaching system design, system framework design based on electric remote control teaching, collection and detection module design, sound identification of hardware design, automatic tracking sensor design, speech recognition software design, software design, realized the electric remote control the design of the teaching system, experimental data show that the proposed intelligent ubiquitous learning under the network model of teaching system design, electric remote control remote control teaching with high precision. Keywords: Ubiquitous intelligence · Learning network model · Electric remote control · Teaching system design

1 The Introduction In the increasingly developed communication technology, information technology, radio frequency identification technology and other new technologies are constantly promoted, a kind of ubiquitous network architecture that can realize direct communication between people, people and machines, people and objects, and even between objects and objects is increasingly clear, and gradually into People’s Daily life. In the U network of ICT convergence technologies, the focus of development has shifted to specific services rather than “technologism”. The goal of pan-network construction is also to provide users with better application and service experience. However, the remote control teaching system realized by traditional technology, especially the remote control teaching system of electrical and mechanical majors with strong operability, has various problems. Therefore, this paper proposes the design of the electric remote control teaching system under the intelligent ubiquitous learning network model. Literature [1] with the high degree of industrialization in modern society, and the continuous progress of science and technology, electrical control automation technology has become the development trend of modern industrial technology. The development of © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 215–226, 2020. https://doi.org/10.1007/978-3-030-63955-6_19

216

Z. Gewei

remote electrical automation system based on embedded technology, This paper mainly studies the design and analysis of remote electrical control automation system based on embedded system. Literature [2] discusses the design and implementation of scheme I from three aspects of training equipment body, upper computer and lower computer, and discusses the design and implementation of scheme II from three aspects of training equipment body, upper computer and lower computer. Literature [3] compares the test, analysis process and actual use effect of various schemes, and launches from three parts: training equipment body, lower computer control part and upper computer simulation software. The training equipment body adopts the modular structure of the combination of patch cord and mesh plate, which is more widely used and more effective. Intelligent ubiquitous learning under the network model of electric remote control architecture is the most critical teaching system design and network technology, the wireless user context in a network environment (environment), perceived performance, heterogeneous wireless access networks coexist with synergy, heterogeneous network mobility management, advanced data management technology (including NID, Profile management, content management, etc.), cross-domain cross layer optimization technique. At the same time, it also puts forward specific Suggestions on standardization. The research on ubiquitous network standard system mainly includes the following contents: the technical characteristics and application objects of ubiquitous network; Ubiquitous network system architecture; Functional modules and components in the system; Interface between modules; Data marks (acquisition, processing, transmission, storage, query, etc.); Application of service standards; Information security, personal privacy protection, etc.

2 Electrical Remote Control Teaching System Design 2.1 Frame Design of Electric Remote Control Teaching System The intelligent remote control teaching system based on the learning network model is a kind of system designed by using the intelligent learning network model. In order to ensure the real-time performance of the teaching system, Virtual Reality Transport Protocols is selected as the system communication protocol module. The Virtual Reality Transport Protocols module is used to realize data communication in multiple geographical locations in the teaching system. At the same time, the Virtual Reality Transport Protocols module is used to enable all the hardware in the teaching system to work together and share resources. In order to ensure the stable operation of the system, run the extension framework module USES hadoop data frame module, architectural patterns adopted B/S structure, core components used the HDFS and graphs, and used the Virtual RealityTransport big data frame module as the communication protocol module system, realized the big data frame module connected to the computer administrator, relying on Internet technology, improve the running speed of the system. Based on the analysis of traditional electronic control remote teaching system, the electronic control teaching system module design is divided into teaching information system management information storage module and information retrieval and analysis module, the main design is divided into two modules. MIS management information

Design of Electrical Remote Control Teaching System

217

storage module, through the large data source, the use of SQL statements, the information in the electronic control teaching system for sampling and analysis, and information storage. The information retrieval and analysis module contains a microprocessor, which converts the information according to the retrieval information instruction input from the water supply end of the system, forms the logical control word structure, analyzes the management information of the teaching information system, and then retrieves the information. Table 1 shows the command word structure with the aid of call analysis logic. Table 1. Retrieves the analysis logic control word structure Serial number

Logical control word structure

Significance

1

A00100101

Transformation

2

0 A 01001

Processing interruption

3

10 A 1001

Internal serial port receiving

4

010 A 100

Synchronous reception

5

0100 A 100F

Peripheral interrupt

6

A 100100H

Generate electrical signal

7

010011H A

External output

8

FF-EFH

I/O output data information

Information analysis module to be obtained by the microprocessor will obtain information instruction and logic control structure transformation, makes the module to analyze data to be obtained, internal DSP buffer are also beginning to action at the same time, the analysis of the received signal is converted into electrical signals to be obtained, to speed up the information analysis module performs data information analysis is taken to be obtained. Through the design of the system communication protocol module, the operation extension framework module and the information system module, the main framework module of the electrical remote control teaching system is designed. 2.2 Design of Data Acquisition and Detection Module for Electric Remote Control Teaching System Teaching system data acquisition and detection module design is the electrical remote control teaching system hardware, two more important hardware modules, is the data acquisition module and data detection module. The data acquisition module of the electrical remote control teaching system contacts directly with the electrical equipment to capture relevant network data. The schematic diagram of the data acquisition module of the electrical remote control teaching system is shown in Fig. 1:

218

Z. Gewei

Link layer driver

Copy data

Buffer

Buffer

Acquisition module

Forwarding data Protocol analysis module Fig. 1. Schematic diagram of data acquisition module of electrical remote control teaching system

The execution power of the data acquisition module of the electrical remote control teaching system can be set according to formula (1): R=

vDP fLd

(1)

Where, D stands for voltage coefficient of electrical equipment; v represents current coefficient of electrical equipment; F represents the operating safety factor set; L d represents the environmental coefficient of electrical equipment. According to the set power of the data acquisition module of the electrical remote control teaching system, the buffer size is calculated [4], as shown in formula (2): Q=

f · D(S − x) × R ρ

(2)

Where, R stands for set execution power, unit m; S stands for copy rate, per bit/S; ρrepresents the current network bandwidth in MB; X is the delay effect factor. The data acquisition module of the electrical remote control teaching system is based on the link layer driver to copy the acquired data and send it to the protocol analysis module. After the protocol analysis module is separated, the separated data is sent to the data detection module design. The detection of network data by data detection module is realized by means of rule file, rule loading module and database. 2.3 Speech Recognition Hardware Design of Electrical Remote Control Teaching System In order to make the electric control distance teaching system can run in the complex electrical equipment. Microcontroller Unit; MCU is composed of speech acquisition,

Design of Electrical Remote Control Teaching System

219

speech processing, feature extraction, speech recognition and other modules, and the Field Programmable Gate Array is divided into each module through FPGA (fieldprogrammable Gate Array). Provides a protocol stack for data transfer between each module. Besides, in addition to teaching, the Internet of things technology is applied to arrange the sound sensor, connect the sound sensor with A/D analog-digital conversion chip, and convert the collected analog signals into digital signals. The selected A/D analog-digital conversion chip should meet the requirements in Table 2, and the parameters given in the following table are the minimum values [5]. The selected A/D ADC chip parameters should be greater than those listed in the table below. Table 2. A/D analog to digital converter chip technical parameters Project

Parameter requirement

Remarks

Conversion rate SPS

1Msps

Number of conversions per second

Resolution bit

12bit

Primary conversion bandwidth

Output interface

SPI

SPI flash

Encapsulation condition

SO patch

More suitable for teaching environment than dip

Taking the converted digital signal as the research object, the converted digital signal is collected by the speech acquisition module, and the speech information is acquired by the speech processing module, feature extraction module and speech recognition module, and stored in the static random access memory (Stram). In addition, in order to ensure good communication effect, quartz crystal resonator is used in the electronic control distance teaching system to reduce interference and other electromagnetic waves. Select the SPI flash communication interface for data communication. Display the display with LED. Press the button to complete the process. The PC side USES the JAVG module to connect. It is the hardware component of a remote electrical control teaching system, as shown in Fig. 2.

Fig. 2. Hardware composition of electrical remote control teaching system

220

Z. Gewei

Physical network sensor Identification technique, is refers to the use of Radio Frequency Identification technology (Radio Frequency Identification, RFID) will sound sensor outside teaching combined with wireless network protocol zigbee technology, short distance transmission at low speed, using infrared technology, bluetooth technology as a whole, at the same time the use of multiple sound sensor to stereoscopic teaching voice signal recognition [6]. Its iot sensor recognition technology is composed of three layers: perception layer, network layer and application layer. The perception layer includes data collection and data short-distance transmission. The data collection is the sound sensor, and the author of the data transmission adopts short-distance transmission. Based on the network, RF identification technology connects adjacent sound sensors for long-distance transmission of data, and transmits the acquired analog signals to A/D analog-to-digital converter chip, which then processes the data in the next step. Figure 3 shows the organization chart of iot sensor identification technology.

Human computer interaction:

Hardware manufacturing technology, etc

Data processing:

Program development, knowledge development Present technology

Long distance data transmission :

Long distance communication technology Network technology, etc

application layer

network layer

Short distance data transmission : RFID, barcode, Bluetooth, infrared, Wi Fi, ZigBee, industrial field bus, etc

Short distance communication and networking

Perceptual level Data collection:

The whole process of information processing

Detection technology, etc Information processing Some key technologies

Fig. 3. Organizational structure of iot sensor identification technology

In order to ensure the validity of the data, the rfid distance should not be greater than the effective distance required by formula (3) [7]. S=

∂[e2 · n] ∂w · n

(3)

Where, e represents the conversion rate of sound sensor, unit Msps; W stands for resolution of sound sensor, unit bit; N is the coefficient of interference. The factors that affect the interference coefficient mainly include the interference speed, the voltage of full load electrical equipment, the interference condition of electrical equipment and so on. Based on the hardware structure design of electrical remote control teaching system and the introduction of sensor recognition technology of Internet of things, the speech recognition hardware design of electrical remote control teaching system is completed.

Design of Electrical Remote Control Teaching System

221

2.4 Design of Automatic Tracking Sensor for Electric Remote Control Teaching System The electric remote automatic tracking is very necessary in the electric remote control teaching system, which is convenient for teachers to solve the trouble of repeated adjustment. Self-tracking sensor a device used to identify a tracking target and determine whether obstacles exist in the tracking path. Designed to take into account the teaching system of automatic tracking features, on the basis of conventional sensors increase the SDI - MAHRS device, SDI - MAHRS device using ultrasonic interference and diffraction principle, through the device of the single crystal of launch ultrasonic wave, using the ultrasonic wave interference and diffraction signal processing results, identify the target tracking objects with their own actual distance and the size of the obstacle exists in the path [8]. Add the sdi-mahrs device structure diagram, as shown in Fig. 4. 5

6

4

7

3 1

2

Fig. 4. Schematic diagram of sdi-mahrs device structure

Among them, 1 represents the shell of the sdi-mahrs device, 2 represents the circuit board, 3 and 4 represent the ultrasonic interference and diffraction image processor, respectively, 5, 6 and 7 represent the u-blox neo-6 m module, u-blox neo-10 m module and u-blox neo-20 m module, respectively. At the same time, in the design process of the automatic tracking sensor, it is also necessary to ensure that the various signals sensed can be converted into signals that can be easily measured, and the corresponding signals can be input into the autonomous control module, which will issue instructions to control the automatic tracking system. Since the current is easy to transmit and measure, the current is used as the output of the automatic tracking sensor. Through overall structure design of automatic tracking system, based on ultrasonic interference and diffraction characteristics, complete the automatic tracking sensor design, automatic tracking target identification and route planning is interference route automatic tracking system is the core of the teaching system, automatic identification and tracking route planning is also a kind of serial port protocol process design, process design by tracking, process design to realize the science reasonable path planning. Considering the stability of the communication among the environment information acquisition module, the autonomous control module and the motion performance module, the method of Socket streaming Socket is used to design the target recognition and route planning process. As a transmission control protocol, streaming socket can

222

Z. Gewei

ensure reliable connection-oriented transmission and guarantee the orderliness and nonrepeatability of target identification and route planning [9]. The flow chart of specific target identification and route planning is shown in Fig. 5.

Soceket instantiation

Bind port

Listen monitoring Y target recognition

Is the connection successful

N

Y Create trace path Connect connection Receive

Send

Close

Fig. 5. Flow chart of automatic tracking target recognition and route planning

2.5 Speech Recognition Software Design of Electrical Remote Control Teaching System In analog A/D analog-to-digital conversion, the high-definition recognition program of teaching voice signal should first add condition sequence at the A/D analog-to-digital conversion position of analog signal, so that when analog signal is converted into digital signal, it has a unique time sequence sign, and set the time sequence of recognizing current digital signal as t, and then the high-definition recognition program of teaching voice signal aims at the time sequence The three digital signals of T-1, t and T + 1 in the column are reorganized to form a complete voice command. The process diagram of the teaching voice signal recognition program is shown in Fig. 6.

External memory controller

TIC DMA BUS Processor core

Embedded RAM and its controller Clock and reset

External memory Speech input

Clock input Reset input

SOPC

Hardware acceleration control and IP

Control output

Fig. 6. Process diagram of teaching voice signal recognition program

Design of Electrical Remote Control Teaching System

223

When the teaching interference is accompanied by a large electromagnetic wave, the analog signal obtained by the sound sensor will appear a certain degree of noise. After a/D analog-to-digital conversion, the noise will form a large number of interference signals. In order to reduce the impact of the environment, the converted digital signal is processed for noise reduction, and the noise reduction principle is shown in Fig. 7 (a, b). Clutter Clutter

peak

Trough

a

wave height

Clutter

Clutter (b)

attenuation

c

(d)

Fig. 7. Schematic diagram of noise reduction processing of teaching speech signal recognition

In the above figure, (a) represents the normal acoustic digital signal. Affected by the surrounding environment, the noise is generated at the inflection point of the normal acoustic signal, as shown in figure (b). Figure (c) represents the original analog acoustic signal. Due to the influence of the propagation distance and the surrounding environment, the original model acoustic wave may have the phenomenon of attenuation, which is shown in figure (d). For the convenience of calculation, it is assumed that the original sound wave does not produce attenuation. In order to reduce the generation of clutter, the multi-a/D conversion is controlled, and the control function is shown in formula (4).   S ·O + z dx (4) f (x) = lim t→∞ t · q Where, O represents the measured acoustic impedance of the environment, in PA s/m3; Q represents the measured propagation speed, in M/S; t represents the measured pressure, in KP. According to the clutter suppression function, the attenuation of the original wave is calculated in reverse. Because of the more clutter, it is greatly affected by the surrounding environment. It shows that the greater the attenuation degree of the original wave is, the attenuation degree is calculated by using the reverse derivation, and the compensation calculation of the sound wave is carried out according to the attenuation degree, as shown in formula (5). w·η f (x) (5) w(x) = lim t→∞ n · j In the formula, η stands for signal transmission frequency, unit: kHz; J stands for signal transmission distance, unit: m; depending on the suppression of clutter and the compensation of attenuation, the noise reduction design of electrical remote control teaching system is realized and the speech recognition of electrical remote control teaching system is completed.

224

Z. Gewei

2.6 Implementation Software Design of Electrical Remote Control Teaching System The software design of the electrical remote control teaching system is based on the hardware design. It uses the data source identification module, data acquisition module, protocol analysis module, data detection module, rule loading module parameters to determine the rule file, module communication mode and realize the design process of the electrical remote control teaching system operation. Based on the software design, the user executes the application program, controls the electrical remote control teaching system, sets the BPF buffer based on the web console, which is used to store the execution command, and realizes the user’s control of the electrical remote control teaching system by calling out the execution mode and using the safety filter on the basis of the protocol stack [10], whose user controls the electrical remote control teaching system Schematic diagram, as shown in Fig. 8: application program

application program

...

cache

cache

cache

Filter

Filter

user kernel

Filter BPF

Protocol stack

user

Network layer driver

Network layer driver

Network layer driver network

Fig. 8. Process diagram of user controlled electrical remote control teaching system

In Fig. 8, the protocol stack is a set of communication protocols for each module communication. In order to be compatible with the data source identification module, data acquisition module, protocol analysis module, data detection module and rule loading module, the protocol stack uses single ICP/IP protocol for communication. At the same time, in order to prevent the intrusion data from hijacking the detection system, load the data filter, and improve the performance of the electrical remote control teaching system Operation safety. The implementation software design of electrical remote control teaching system is realized.

3 Simulation Test Experiment In order to ensure the effectiveness of the design of the electrical remote control teaching system (hereinafter referred to as the remote control teaching system) under the intelligent ubiquitous learning network model, the simulation test and analysis are carried out. In the analysis process, the conventional remote control teaching system and the remote control teaching system under the conventional learning model are used as the experimental comparison objects to verify the accuracy of the teaching remote control.

Design of Electrical Remote Control Teaching System

225

3.1 Experiment Preparation In the experiment, the past teaching data is used as the experimental object to carry out the simulation experiment, and the past teaching data is used as the experimental object, including the known teaching content, to analyze 12.5%, 25.0%, 37.5%, 50.0%, 62.5%, 75.0%, 87.5% of the teaching data process. 3.2 Experiment Process This experiment uses different teaching data methods to analyze and compare the completed teaching process, and uses the past parameters to verify the accuracy of the method. Therefore, it is necessary to build a long-distance control teaching system, longdistance control teaching system, long-distance control learning mode, long-distance teaching mode that does not contact the teaching results, etc. suitable for the previous experimental environment. In the process of the experiment, firstly, the experimental environment is established, which is basically in line with the development of the actual situation. The time function is used to control the development of the situation. 3.3 Analysis of Experimental Results According to the time process, the remote control teaching system, the conventional remote control teaching system and the remote control teaching system under the conventional learning model are obtained. According to the situation prediction in different time periods and the recorded data, the experimental results data table is formed, as shown in Table 3: Table 3. Comparison of experimental results Stage of teaching facts

Remote control accuracy Remote control teaching system

Conventional remote Distance control teaching control teaching system system based on conventional learning model

12.5%

60.5%

80.4%

50.4%

25.0%

66.7%

60.5%

48.6%

37.5%

69.8%

54.1%

45.2%

50.0%

70.0%

48.7%

48.5%

62.5%

75.4%

62.2%

54.5%

75.0%

76.4%

68.2%

54.8%

87.5%

87.5%

79.8%

68.7%

From the experimental results, it can be seen that the traditional remote control teaching system has a high sensitivity in the initial stage of teaching facts, but in general,

226

Z. Gewei

the system is greatly influenced by the stage of teaching data, and as more information is received, it is easy to cause errors in judgment. In the traditional teaching mode, the remote control teaching system has high stability, but the overall teaching progress is slightly lower than the existing remote control teaching system. Through the statistics and calculation of the experimental data, the remote control accuracy of the remote teaching system reached 61.56%, the remote control accuracy of the traditional remote teaching system reached 55.96%, and the remote control accuracy of the traditional learning model reached 55.96%45.17%. Remote control teaching system is more effective than traditional remote control teaching system.

4 Conclusion Based on the electric control of the remote teaching system framework design, acquisition and the design of the detection module, voice recognition, automatic tracking sensors, voice recognition software, hardware implementation of software design, realized the electronic control of the remote teaching system design, in order to guarantee the effectiveness of the design system, through the simulation test of experimental data show that the proposed intelligent flood in the design of the remote control system of teaching under the network model has higher accuracy and the remote control of teaching the design has good application prospect. In order to provide a certain theoretical basis for the design of electronic control distance teaching system.

References 1. Postnova, E.A.: Optimal motion control of the system modeled by double integrator of fractional order. Autom. Remote Control 80(4), 761–772 (2019) 2. Bazhenov, S.M., Vakhonina, S.A., Tarasov, N.V., et al.: Information and control system for automated weaving process monitoring. Autom. Remote Control 80(3), 576–583 (2019) 3. Martinov, G.M., Nikishechkin, P.A., Grigoriev, A.S., et al.: Organizing interaction of basic components in the CNC system AxiOMA control for integrating new technologies and solutions. Autom. Remote Control 80(3), 584–591 (2019) 4. Shao, P., Tian, J., Yang, F., et al.: Identification and regulation of active sites on nanodiamonds: establishing a highly efficient catalytic system for oxidation of organic contaminants. Adv. Funct. Mater. 28(13), 1705295 (2018) 5. Robinson, A.E., Lowe, J.E., Koh, E.I., et al.: Uropathogenic enterobacteria use the yersiniabactin metallophore system to acquire nickel. J. Biol. Chem. 293(39), 118–123 (2018) 6. Sullivan, D.R., Ganzini, L., Lapidus, J.A., et al.: Improvements in hospice utilization among patients with advanced-stage lung cancer in an integrated health care system. Cancer 124(3), 14–20 (2018) 7. Shi, X., Yang, C., Xie, W., et al.: Anti-drone system with multiple surveillance technologies: architecture, implementation, and challenges. IEEE Commun. Mag. 56(4), 68–74 (2018) 8. Hiroshi, K., Auld, D.S.: Microanalytical system for determination of picogram quantities of metals in metalloenzymes, as illustrated with zinc-containing enzymes. Clin. Chem. 12(4), 4–10 (2018) 9. Zhao, X., Yang, Y., Wang, Y., et al.: Study of the converter based on photonic crystals filters and quantum dots for solar blind ultraviolet imaging system. Opt. Eng. 57(9), 45–51 (2018)

Research on the Integrated Mode of Ideological and Political Education in Colleges and Universities Based on Multivariate Data Analysis Zhu-zhu Li and Ming-jun Cen(B) Nanning University, Nanning 530200, China [email protected]

Abstract. At present, ideological and political education is highly valued by the state. In order to improve the level of ideological and political education in universities and accelerate the popularization of ideological and political education, an integrated model of ideological and political education in universities based on multivariate data analysis is proposed. Taking the development of ideological and political education as the basic goal, while analysing the inevitable requirements of the integration model, the main classification criteria of the ideological and political education research objects are clarified, and the research on the education integration model is completed. College education. On this basis, the principle of integration processing is proposed, and various data attributes of ideological and political education in colleges and universities are defined. Then, combined with the mapping conditions under the integration mode, a comprehensive environment of ideological and political education based on multivariate data analysis is established. With the aim of grafting the ideological and political education model, we explore the development methods of ideological and political education in colleges and universities, determine the implementation law of ideological and political education for continuous and harmonious development, and propose constructive methods to develop the integrated model of ideological and political education in colleges and universities. Keywords: Multivariate data · Ideological and political education · Mapping condition · Mode grafting

1 Preface The traditional theoretical knowledge of data statistics is the foundation of multivariate data analysis. Therefore, multivariate data analysis can be regarded as a branch of traditional data statistics and is a comprehensive analysis method. When multiple objects and indicators are related to each other, statistical rules for each object are obtained through analysis. Multivariate data analysis is applicable to ideological and political © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 227–238, 2020. https://doi.org/10.1007/978-3-030-63955-6_20

228

Z. Li and M. Cen

fields, including multivariate normal distribution and its sampling distribution, multivariate normal population mean vector and covariance matrix hypothesis test, multivariate analysis of variance, linear regression and correlation, multiple linear regression And correlation (I) and (II), main component analysis and factor analysis, discriminant analysis and cluster analysis, Shannon information content and its application [1]. This is called multivariate analysis. When the distribution of the population is a multidimensional (multivariate) probability distribution, please use mathematical statistical theory and methods to deal with the population. According to the research characteristics of the ideological and political integration model, multiple data analysis can be divided into multiple basic types of multiple regression analysis, discriminant analysis, cluster analysis, principal component analysis, correspondence analysis, factor analysis and typical correlation analysis [2]. Ideological and political education is the foundation of social progress and development. Ideological and political education can improve the political opinions and moral restraint of the masses, so as to be a bit face-to-face, from the individual to the collective, to improve the ideological and political level of the entire society, so that they can form a social ideology And ethical social activities. A society. This is also a lifelong learning course in Chinese literature. Ideological and political education is the main content of China’s spiritual civilization construction and one of the main ways to solve social conflicts and problems. Although ideological and political education is important, it is difficult to achieve universalization. Especially in the market economy, the ability of ideological and political education needs to be improved and cannot meet the needs of social development. It is because of our neglect of ideological and political education [3– 5]. Personality education is the foundation of the application of ideological and political education. Without this framework, ideological and political education is like a rootless duckweed. Aiming at the problem of the limited development of the integration model of ideological and political education in colleges and universities, a multi-data analysis method is proposed. In studying various political views and ethics, he also pointed out the current status of malfeasance in the current development situation and put forward a series of instructions. Rectification suggestions (Fig. 1).

2 Integrated Model of Ideological and Political Education in Colleges Ideological and political education is the main development direction of the university integration model, and its main research content can be summarized into three basic directions: educational connotation, necessary requirements, and the target of education. 2.1 Connotation of Ideological and Political Education Development Ideological and political education is a new direction for the development of quality education in colleges and universities. To some extent, it is different from traditional colleges and universities’ ideological education, but it actively serves all tasks related to students. Summarizing the development of ideological and political education in universities, this application mode has the following development characteristics. First, the development

Research on the Integrated Mode of Ideological

229

Fig. 1. Development needs of ideological and political education in universities

of ideological and political education in colleges and universities has a strong breakthrough. The improvement of integrated technical level breaks through the limitations of traditional ideological and political education and provides more favorable conditions for the in-depth development of ideological and political education in colleges and universities. Compared with the limitations of traditional behavioral education, ideological and political education in colleges and universities has better broken the role restrictions between educators and college students, paid more attention to equality between the two, and thus more respected the subject status of the educated. Make it better for interactive communication. The educating and guiding function of online ideological and political education in colleges and universities has benefited more [6]. Secondly, colleges and universities are the places where the density of students is the greatest. College students can not only learn from vast behavioral data, expand their horizons, and gain insights, but also relax and entertain themselves physically and mentally. Therefore, they are highly sought after by college students. Finally, the guiding role of the development of ideological and political education in universities is obvious. The integrated education information is complex and diverse, the level is uneven, and the negative influence of disseminated negative speech on ideological and political education in colleges and universities must not be underestimated. It is precisely under such a development model that the development of ideological and political education in colleges and universities has always faced severe challenges. Applied issues. 2.2 Inevitable Requirements of the Integration Model In the context of ideological and political education in colleges and universities, the integration model and the basic education system must be effectively connected in order to exert the most direct effect. Compared with other educational organizations, although ideological and political education in universities has been in a relatively rapid development mode, it is still in its infancy, and a relatively complete research system has not yet been established for this type of development mode. Personality is the basis for the

230

Z. Li and M. Cen

stable formation of the outlook on life and values. Only when a person’s three outlooks are unified can they have a stable value concept and a stable psychological basis for life values. Human values must be unified and stable, which requires a person’s psychological process and personality form to be unified and stable. Otherwise, split personality can only produce the concept of split [7]. Personality is the inner psychological basis for the formation of a specific world outlook and outlook on life. The world view is an understanding of the world. Although the correct world view comes from the correct theoretical guidance and learning, if there is no benign personality form as the internal psychological basis, external instillation will be difficult to play a role. The integrated model is the main driving force for the formation of specific moral qualities. Personality has qualitative characteristics. Therefore, once a personality is formed, a person will have a corresponding internal texture. Different textures will adapt to different moral tendencies, and a benign personality is easy to establish a benign moral quality. Of course, these basic roles of personality are not absolute, but relative. At the same time, they also interact and change with people’s values, world outlook, life outlook and moral consciousness [8]. Therefore, personality has both unity and stability, as well as differences and variability. These characteristics also determine the difficulty of developing a benign personality. In short, the personality state can be said to be the subtle, concealed and primitive state of ideology and morality, while the ideology and morality are often the development, shaping, maturity and clear personality expression (Table 1). Table 1. Conditions for the establishment of an integrated model of ideological and political education in universities Foundation condition

Necessary establishment condition

Auxiliary conditions

Thought-level condition

Personality

Theoretical guidance

Moral values

Behavior motivation

Mental process

The theory of learning

Emotional concept

Ideological motivation

World view

Psychological basis

Moral tendency

The moral force

The outlook on life

Behavior on the basis of

Psychological basis

Emotional expression

Moral consciousness

Thoughts on the basis of

The value of life

Behavioral expression

2.3 Research Objects of Ideological and Political Education in Colleges and Universities Ideological and political pedagogy is a discipline that guides people to form correct ideological behaviors. It takes the law of the formation and change of human ideological

Research on the Integrated Mode of Ideological

231

behaviors and the law of ideological and political education as its own research objects. Among them, the change of people’s thoughts, viewpoints and positions, as well as the formation of the outlook on life and the world, are the focus of research. Ideological and political pedagogy is a complex network system. With the promotion of the integration model, its main research objects are: First, study the physical and psychological factors of human beings, the research needs — motivation-the development process of behavior, and reveal the development law of human’s own thoughts and behaviors. Secondly, to study the relationship between the general relationship, the complex social factors and the formation and development of the three views of the education object, and to analyze how to adjust the impact of the social environment on people across dimensions. Third, study the management system and leadership function of ideological and political education, evaluate the quality of ideological and political education workers, and explore the path of ideological and political education penetration into the commercial field, so as to promote cooperation in various fields and achieve the purpose of collaborative education. The law of human thought activity. There is a mutual influence between human thoughts and behaviors and their living environment. The social environment produces human needs, and the demand for human motivation also determines human behavior. Some behavioral results are fed back to regulate demand, increase motivation and re-dominate human behavior [9]. This periodic movement reflecting the interaction and mutual restriction of various ideological factors constitutes the process of people’s ideological activities.

3 Integrated Ideological and Political Education Environment Based on Multivariate Data Analysis With the support of the integrated model of ideological and political education in colleges and universities, according to the operating procedures established by integrated processing, multiple data definitions, and integrated mapping conditions, an integrated ideological and political education environment based on multivariate data analysis is established. 3.1 Integrated Processing of College Ideological and Political Data Parallel coordinates of multivariate data is one of the mainstream technologies for the integration of ideological and political information. A significant advantage of parallel coordinates is that it has a good model foundation, and its projective geometric interpretation and duality characteristics allow it to express the projection information of high-dimensional data in multiple low-dimensional subspaces simultaneously [10]. But the integration of parallel coordinates mainly focuses on the visualization of the original ideological and political data, not on the specific university data analysis task. Because a point-line transformation is used in parallel coordinates, it is easy to produce view aliasing, that is, over-drawing problems. This problem is especially serious when the number of samples in the ideological data set is large. When the goal of integrated data

232

Z. Li and M. Cen

analysis is pattern recognition, the most concerned thing is not all the information of each sample itself, but the difference between classes. That is to say, it is not necessary to visualize all the integrated data samples, and only the ideological and political samples with a special contribution to the classification can be visualized. For multiple linearly separable cases, the integrated vector set is a subset of the ideological data vector. The integration vector refers to all the ideological and political education samples located on the extreme edge of the training set, that is, the samples located on the convex hull of the training set. In fact, the integration vector is the convex vertex of the sample set. Therefore, using the integrated processing algorithm to crop the ideological and political data training set samples and using the integrated vector of the training set as the representative of the training set to draw parallel coordinates can reduce the view aliasing problem of parallel coordinates and facilitate the observation of ideological and political integration Data structure and feature information to distinguish multiple categories. Set p¨ to represent the highest-level disposal conditions for the integration of ideological and political education, p to represent the lowest-level disposal conditions for the integra··

tion of ideological and political education, and W to represent the integrated order of magnitude of multiple ideological and political data samples. The integrated processing results of the data are expressed as:    p¨     (W − χ y1 )2 dy1     p··  e=  (1) p¨   βy2 · ||r1 + r2 ||dy2 p ··

Among them, χ , β, represent the conditions for the application of multiple data and integrated analysis in the context of ideological and political education in colleges, y1 , y2 , represent the data samples of two different ideological and political education in colleges, and r1 , r2 , represent integrated ideological and political Multivariate classification coefficients and characteristic information parameters of the data. 3.2 Definition of Diversified Data of Ideological and Political Education in Colleges and Universities The hierarchical structure of ideological and political education data in colleges divides the diversified space into rectangular subspaces, and the size of the subspace is determined by the size of the integration node. The level of the tree map is based on the order from the root node to the leaf node, and the horizontal and vertical conversions are sequentially performed. The rectangular subspace is divided horizontally. The next layer will be divided vertically, and the next layer will be divided horizontally. analogy. For each divided ideological and political data collection space, corresponding pattern matching or necessary explanation can be performed. For the integration of multivariate data, there is very little data diversification mapping method that can directly map the data information and the information that the model subject wants to know. Generally,

Research on the Integrated Mode of Ideological

233

multivariate methods are used to deal with this mapping relationship. In order to better understand the diversity of ideological and political education data, the process of understanding information can be layered. Hierarchy is generally divided into two steps. The first step is to understand the approximate distribution information of the data through a multiple view, such as density, pattern, and size. You can understand the general distribution range or structure information of the information; the second step is to observe the details with another view on the selected detailed data area. In order to explore the data relationship between the various data dimensions in a multi-dimensional data set, two-dimensional data can be represented in a scattered manner; but the more the data has more dimensions, the more difficult it is to process. As the number of dimensions increases, the number of possible diversified portfolio index levels continues to increase. 

Set I to represent the upper-level grading conditions for the data of multiple ideological 

and political education, and I to represent the lower-level grading conditions for the data of multiple ideological and political education. The simultaneous formula (1) can define the diverse data of ideological and political education in colleges as:     I  (q1 + q2 )2 e   I (2) R= 3u |T |/χ E In the above formula, q1 and q2 respectively represent two different data parameters of ideological and political education in colleges and universities, u represents the application information of the integrated model, T represents the cyclical implementation time of ideological and political education, and χ represents the modality of ideological and political education Analysis conditions, E represents the number of ideological and political samples in an integrated education environment. 3.3 Mapping Conditions for Integration of Ideological and Political Education As the main representative of the subspace coordinate integration method, the projective geometric interpretation of parallel coordinates has great potential for educational pattern recognition tasks, but the problem of multivariate data analysis hinders its practical application in the field of integrated analysis. The main reason for the parallel coordinate integration is that its point-line dual transformation increases the average data proportion of data information. This point-to-point physical mapping relationship integrates parallel coordinates and scattered point coordinate systems into the same integrated framework. This multivariate data set is equivalent to integrating dual coordinates and parallel coordinates in the same model environment, so it is called Parallel dual data collection. The same sample has both point and line representations in parallel dual integration. The two forms have a certain geometric correspondence. You can transform as needed to adjust the information parameters between the data. Ratio, to alleviate the problem of the deviation of the integrated model, and provide a powerful tool for the identification of multiple ideological and political education models. In the practice of ideological and political education in colleges and universities, describing an objective data collection

234

Z. Li and M. Cen

object usually requires multiple integrated attribute variables, so the number of information attribute variables is greater than 1. Such data is generally called multivariate data. Because the educational subject has a strong sense of diversity, the integration of ideological and political data is of great significance for understanding and processing multiple data. As we all know, no one type of data information set can be applied to all model environments, so the choice of ideological and political connotation according to the purpose of the education and the specific environment, or the comprehensive use of a variety of diverse data information, can directly make the integration of ideological and political education in colleges The conditions are met. Let f represent the parameter proportion conditions of ideological and political education information in colleges and universities, and ξ1 and ξ 2 respectively represent two different multi-dimensional data collection objects. The simultaneous formula (2) can express the integrated ideological and political education mapping conditions as:     a↑    (f − dl)2 R¯j   a↓ (3) S=  a↑   k˜ · ||ξ1 + ξ2 ||d k˜ a↓

Among them, a ↑ represents the upper limit data analysis conditions in the integrated mapping relationship, a ↓ represents the lower limit data analysis conditions in the integrated mapping relationship, d represents the analysis and disposal authority of multiple ideological and political education data, and ˙l represents the integration of ideological and political education information in universities Behavioral mode performance, ¯j represents the average characteristic coefficient of the integrated information in the multivariate data set, and k˜ represents the quantitative analysis of the ideological and political education data of colleges and universities.

4 Rectification and Modification of the Integrated Model of Ideological and Political Education in Colleges and Universities Combined with the conditions of multivariate data analysis, from the three aspects of ideological and political model grafting, development needs exploration, and implementation of law research, put forward suggestions for the development of university ideological and political education integration model renovation and modification. 4.1 Grafting of Integrated Ideological and Political Model The idea of model grafting is a process of multi-data fusion. Ideological and political education in universities should also accept the idea of knowledge grafting. This is not only the learning philosophy of other disciplines, but also a new requirement for ideological and political education in universities under the integrated research model. In a sense, knowledge also has basic characteristics of life, such as data information metabolism.

Research on the Integrated Mode of Ideological

235

In the process of rectification of multiple data, the ideological and political model grafting mainly involves the teaching of theoretical courses in colleges and universities, that is, the knowledge of ideological and political theory courses that have been processed, specialized, structured, and have “life” are transplanted into the minds of learners The process of going through the original knowledge structure can be regarded as “grafted” when taught; self-learning can be regarded as “self-grafted”; teaching and learning technology is actually knowledge grafting technology. According to the needs of college students and the needs of problem solving, purposeful and targeted knowledge grafting is conducive to the knowledge innovation of ideological and political theory courses. By using this concept to cultivate reliable successors and qualified builders for socialism. For ideological and political education in colleges and universities, personalized learning is based on the individual characteristics and development potentials of learners, and adopts a suitable way to fully meet the individual needs of learners. The integrated model brings a huge amount of information resources. Students can obtain different information and knowledge according to their own needs. Although the manifestations of these information are different, this expands, expresses and elaborates the internal diversity of knowledge to a certain extent. Conducive to the individual needs of learners with different learning abilities and styles. At the same time, the multi-education transmission mode brought by multi-data analysis can also teach students according to their aptitude. In this way, students can freely choose a virtual or hybrid multi-learning environment, which greatly improves the learning efficiency (Fig. 2).

Fig. 2. Grafting principle of integrated ideological and political model

236

Z. Li and M. Cen

4.2 Development Needs of Ideological and Political Education in Colleges Ideological and political education is a discipline that is constantly developed and updated in practice. Advances in multivariate data analysis have created a new horizon for traditional ideological and political education, accumulated rich resources for the long-term development of ideological and political education in colleges, and urgently required college education organizations to strengthen theoretical and practical research in order to further enrich ideological and political education The theoretical system guides the practice of ideological and political education with high-level theoretical research results. In order to meet the needs of the times and social development, network ideological and political education in colleges and universities must continuously promote the innovation of education forms, that is, systematic, timely, active, diversified, and interactive. Systematic means that the construction of university ideological and political education should rely on an integrated development platform, comprehensively utilize ideological and political education materials, create new methods of ideological and political education, open up a new situation in ideological and political education, and then achieve comprehensive education, humanistic education and ideological and political education. Organic integration of education; Timeliness means that the ideological and political content needed for education needs to be updated in a timely manner, and subjective issues must be reflected in a timely manner; Activeness means that the educator always maintains a dominant position and promptly and proactively judges the personal education needs, or according to the specific needs of college students Make judgments and respond quickly to the situation; diversification means that the content and forms of online ideological and political education in colleges and universities must be diversified, making full use of multiple implementation modes such as integrated platforms, and multi-pronged approaches such as language, writing, communication, and influence. Students are more receptive; interaction refers to the interaction between educators and student groups, to exchange ideas through multiple information media, and to solve practical application problems (Table 2). 4.3 Adhere to the Rules of Ideological and Political Education for Harmonious Development Ideological and political education is an important work to solve ideological problems, it is the spiritual support for the progress of the social integration model, and it is a strong guarantee for the training and transportation of qualified personnel for the society. It is required to be consistent with the harmonious development of society. As a field of ideological and political education in colleges and universities, ideological and political education also needs harmonious development. Whether it is the construction of a theoretical system, or educational content, work methods, etc., it needs harmonious development. Important support must always adhere to the guidance of Marxist principles, apply scientific worldviews and methodologies in practice, and strive to promote the construction of socialist ideological positions in a multi-data analysis environment. Purposeful indoctrination and guidance are required to allow advanced Educational philosophy is fully integrated with mainstream ideology. Ideological and political education in colleges and universities should adhere to the guidance of important ideas such as

Research on the Integrated Mode of Ideological

237

Table 2. Basic development needs of ideological and political education in universities Ideological and political education connotation

Theoretical system teaching resources

A system of ideological and political theories

Ideological and political education practice

Ideological and political education development behavior

Integrated teaching model

Guiding ideological and political work

Ideological and political teaching needs

Characteristics of ideological and political education development approach

Systematic

In time

The initiative

Educational connotation interpretation

Relying on the integrated development platform

Timely update ideological and political education content

Ideological and political education needs

Marxism-Leninism, student-oriented, advancing with the times, and harmonious development, closely integrate the actual situation of ideological and political education in colleges and universities, and strive to realize the theory and content of ideological and political education All-round innovation in aspects, methods, mechanisms, etc., adhere to the integration of features, follow the rules of students’ ideological development and meet the actual needs of students, unify the three, strengthen the guidance of education with the help of a multi-information platform, and build a new stage of “thinking and cultivating talents” By spreading mainstream culture and ideas through model media, it will greatly enhance the sense of the times and influence of ideological and political education in colleges and universities, and train national successors with high comprehensive quality.

5 Conclusion This paper cites multiple data analysis methods to solve the current problem of integration of ideological and political education in colleges and universities. After clarifying the development direction of the ideological and political education and the inevitable requirements of the integration mode, the research of this article is completed through the integration of the model. This article establishes a good ideological and political education environment based on multivariate data analysis. By exploring the future development direction of ideological and political education, the implementation method of ideological and political education is clarified and the integrated teaching model is determined. Based on the implementation of the harmonious development of the ideological and political education implementation rules, it provides new ideas for the future development of ideological and political education.

238

Z. Li and M. Cen

References 1. Shanyong, L.: A new fracability evaluation approach for shale reservoirs based on multivariate analysis: a case study in Zhaotong Shale Gas Demonstration Zone in Sichuan. China. ACTA GEOLOGICA SINICA (English edition) 93(4), 1005–1014 (2019) 2. Tian, Y., Gou, X., Niu, P., Sun, L., Guo, Y.: Multivariate data analysis of the physicochemical and phenolic properties of not from concentrate apple juices to explore the alternative cultivars in juice production. Food Anal. Methods 11(6), 1735–1747 (2018). https://doi.org/10.1007/ s12161-018-1169-2 3. Wang, H., Mei, C., Liu, J.H., Shao, W.W.: A new strategy for integrated urban water management in China: sponge city. Sci. China Technol. Sci. 61(3), 317–329 (2018). https://doi. org/10.1007/s11431-017-9170-5 4. Patel, R., Cox, R., Correll, N.: Integrated proximity, contact and force sensing using elastomerembedded commodity proximity sensors. Auton. Robots 42(7), 1443–1458 (2018). https:// doi.org/10.1007/s10514-018-9751-4 5. Chaudhari, K., Ukil, A., Kandasamy, N., et al.: Hybrid optimization for economic deployment of ESS in PV-integrated EV charging stations. IEEE Trans. Ind. Inform. 14(1), 106–116 (2018) 6. Deans, H., Mirjam, A.F., Ros-Tonen, M.D.: Advanced value chain collaboration in ghana’s Cocoa Sector: an entry point for integrated landscape approaches? Environ. Manage. 62(1), 143–156 (2018) 7. Ahmadivand, A., Gerislioglu, B., Tomitaka, A., et al.: Extreme sensitive metasensor for targeted biomarkers identification using colloidal nanoparticles-integrated plasmonic unit cells. Biomed. Optics Express 9(2), 373–386 (2018) 8. Tong, X., Sun, H., Luo, X., Zheng, Q.: Distributionally robust chance constrained optimization for economic dispatch in renewable energy integrated systems. J. Glob. Optim. 70(1), 131–158 (2017). https://doi.org/10.1007/s10898-017-0572-3 9. Adly, M.: Kai Strunz: irradiance-adaptive PV module integrated converter for high efficiency and power quality in standalone and DC microgrid applications. IEEE Trans. Indust. Electron. 65(1), 436–446 (2018) 10. Reiskarimian, N., Dinc, T., Jin, Z., et al.: One-way ramp to a two-way highway: integrated magnetic-free nonreciprocal antenna interfaces for full-duplex wireless. IEEE Microwave Mag. 20(2), 56–75 (2019)

Simulation Training Remote Control System of Industrial Robot Based on Deep Learning Dan Zhao and Ming Fei Qu(B) Beijing Polytechnic, College of Mechatronic Engineering, Beijing 100176, China [email protected], [email protected]

Abstract. In order to improve the remote control performance of industrial robot simulation training, deep learning algorithm is used to optimize the design of traditional remote control system. On the basis of traditional remote control system, the configuration of hardware system is modified, and the database of control system is established. With the support of hardware system and database, the remote control of two training items of industrial robot simulation mobile training and simulation picking training are realized respectively. Through the system test experiment, the conclusion is drawn: compared with the traditional industrial robot remote control system, the control function of the design control system is improved, and the system can save about 12.5 s response time in the control process. Keywords: Deep learning · Industrial robot · Simulation training · Remote control system · System design

1 Introduction Industrial robot is a multi joint manipulator or multi degree of freedom robot facing the industrial field. It is a kind of machine which can automatically perform work and realize various functions by its own power and control ability. It can be commanded by human beings or run according to the pre arranged program. Modern industrial robots can also act according to the principle program formulated by artificial intelligence technology [1]. Nowadays, industrial robots can replace people to do some monotonous, frequent and repeated long-term operations, or work in dangerous and harsh environments. Before the industrial robot is put into application, it needs to carry out simulation training. The simulation training can be roughly divided into two parts, namely, the mobile performance training and the picking performance training of the industrial robot [2]. The above-mentioned simulation training projects of the industrial robot need to be realized with the help of the corresponding control system. Industrial robot control system is the brain of robot, which is the main factor to determine the function and performance of robot. Under the influence of information technology, the remote control system of industrial robot simulation training is formed

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 239–251, 2020. https://doi.org/10.1007/978-3-030-63955-6_21

240

D. Zhao and M. F. Qu

by integrating network communication technology and industrial robot control technology. The remote control system is generally divided into two parts: client program and server program. Before use, the client program needs to be installed on the main control computer, and the server program needs to be installed on the controlled computer. Remote control can only be carried out through the network. The local computer is the sending end of the operation instruction, which is called the main control end or the client, and the non local control computer is called the control end or the server. “Remote” is different from “remote”. The main control end and the controlled end can be in the same room of the same LAN, or two or more computers connected to the Internet at any position. The traditional industrial robot simulation training remote control system mostly uses robot vision technology, Internet of things technology or cloud control technology. However, after a long time of application research, it is found that the traditional remote control system can not achieve the control of robot movement and picking at the same time, and there is a time delay phenomenon in the control process. In order to solve the above problems in the traditional remote control system, this paper uses the deep learning algorithm to design the industrial robot simulation training remote control system. Deep learning is a new field in machine learning. Its motivation is to build and simulate neural network of human brain for analysis and learning. It imitates the mechanism of human brain to interpret data, such as image, voice and text. The development of deep learning has greatly promoted the innovation of visual perception technology. The method of extracting features from massive data through deep neural network has also shown strong advantages. This also promotes the research upsurge of artificial intelligence in the field of robot. This paper designs and realizes the remote control system of industrial robot from four aspects: hardware, database, human-computer interface and software function, so as to realize the efficient control of industrial robot. And the development of deep learning is expected to solve the challenge of traditional remote control system.

2 Design of Remote Control Hardware System for Robot Simulation Training The simulation training remote control system of industrial robot is based on client and server architecture. The system is composed of computer control terminal, public network management forwarding server and mobile robot. The computer control terminal controls the mobile robot through the public network management forwarding server. The control principle of the industrial robot simulation training remote control system is: the public network management forwarding server starts and waits for the computer control terminal to connect with the mobile robot terminal. After the computer control terminal starts, it sends the registration request to the server. Similarly, after the mobile robot terminal starts, it sends a registration request to the server [3]. Then the computer control terminal selects the mobile robot to be controlled and sends the control request to the service. After receiving the request, the server establishes a communication connection between the computer control terminal and the mobile robot terminal. Then the mobile robot terminal sends the video taken by the mobile robot to the computer control

Simulation Training Remote Control System of Industrial Robot

241

terminal through the server. The computer control terminal receives the video, displays the video, and sends the control instruction to the mobile robot through the server. According to the above-mentioned remote control principle of industrial robot simulation training, the remote control system of industrial robot is designed and implemented from four aspects of hardware, database, human-computer interface and software function, so as to achieve efficient control of industrial robot. The hardware system of industrial robot simulation training remote control system provides hardware support for the realization of control function and the compilation of related programs. The specific hardware system structure is shown in Fig. 1. A serial port

A serial port

The network interface External power supply

JTAG debugging interface ARM minimum system

Flash

SDRAM

Extension interface

Fig. 1. Basic architecture of remote control hardware system

2.1 Industrial Robot The industrial robot terminal is mainly responsible for collecting video and sending it to the transponder, and receiving the mobile control instructions from the transponder. According to the established mobile control protocol, the robot makes the corresponding mobile control motion. In the design of the remote control system, the Toshiba four axis thl400 robot is used. The robot is a fast plane grabbing and placing robot commonly used in the industrial field. The binocular robot assembly system adopts ABB six axis industrial robot. Six axis robot is also widely used in automatic sorting, welding, spraying and other fields in the industrial field. The two kinds of robots use pneumatic suction nozzle to grab the workpiece. The industrial computer and robot controller are connected through Ethernet and communicate through socket protocol. 2.2 Remote Wireless Communication Equipment 2.2.1 Wireless Communication Network Environment Wireless local area network is needed between the simulation training control server and the application server of industrial robot, so that the industrial robot can get a larger range of movement. Select two wireless network cards with USB interface, and install them on the USB port of host a and host B respectively, then install the driver of wireless network card and supporting application program. Just make sure that A and B can be connected, and only use the utility provided by the network card to set up the WLAN. When you are finished, you can view the connection instructions to the system.

242

D. Zhao and M. F. Qu

2.2.2 The Server In the remote wireless communication network environment, the servers that need to be accessed include Internet server and robot server. The Internet server consists of three parts: Web service program, user database and Winsock communication program. The Internet server of the remote robot is a server with a static IP address. It runs the Apache 3.12 HTTP service program under Windows NT 4.0. When the user connects to the operation home page of the remote robot through a web browser, he starts the web service program of the Internet server [4]. The web service program provides users with static pictures and robot parameters of the operation site by operating the HTTP file of the home page. Users manipulate the PUMA robot on the site by changing the robot parameters. Robot server is an industrial computer running Windows 2000, which is equipped with robot communication control card, image acquisition card, gripper controller and other hardware. In the remote control system, the main work is to modify the robot’s motion in the tool coordinate system and read the transfer matrix in the end world coordinate system. 2.3 Remote Control Equipment The remote control of industrial robot is composed of two five phase stepping motors and corresponding motor drivers. Two stepper motors drive two driving wheels. By changing the frequency of the pulse signal acting on the stepper motor controller, the stepper motor can be adjusted with high precision. At the same time, applying the same or different pulse signals to the two stepping motors can control the motion of the industrial robot conveniently. 2.4 Circuit Design In addition to the remote control equipment and remote communication equipment of the above remote control system, the hardware system also includes the minimum system, debugging interface, network interface, external power supply, flash memory, SDRAM and expansion interface. Among them, the minimum system refers to the minimum configuration system in which the microprocessor can run programs and complete the simplest tasks. The minimum system is a necessary part of any complex system. The minimum system includes: power circuit, crystal oscillator circuit, reset circuit, etc., debugging test circuit, etc. [5]. The power circuit is used to supply power to each module to ensure its normal operation. The power supply of the system is 5 V, the power supply voltage of the chip pin is 3.3 V, and the power supply of WiFi module is 5 V, so necessary power conversion is needed. The design circuit of the system power supply circuit is as follows (Fig. 2). In addition, the reset circuit will initialize the processor state to a state that can make it run normally, so that the microcontroller can start to work from that specific state every time it is powered on. This reset logic requires a signal to work properly. The function of crystal oscillator circuit is to provide power for the system. Almost all microcontrollers are sequential circuits, which requires a clock pulse signal to make them work normally. The general microcontroller itself has a crystal oscillator. But some special occasions need to use external oscillator to provide clock signal.

Simulation Training Remote Control System of Industrial Robot U2

5V

3.3V Vout

Vin GN D

C2 0.1uF

243

C2 0.1uF

GN D

Fig. 2. Power circuit connection diagram

3 Design of Remote Control System Database for Robot Simulation Training The database of industrial robot simulation training remote control system is mainly used for user management and remote control data storage. So the design of database includes two parts: user table and basic operation data table of industrial robot. Among them, the user table is used to manage the users logged in by the client, which is mainly implemented in two aspects: entity contact diagram, attribute and database logic design [6]. The database uses MySQL database to create remote control system user table and user authority table respectively. Ordinary users can only view the video of remote control system without management function. Administrator users can manage the system video, including adding, deleting, modifying and querying. Add an identification field to the user table to classify the user, and identify whether the user is a different user or an administrator user. The table structure of the user table is shown in Table 1. Table 1. User database of remote control system Database parameters

Parameter definition

The data type

Id

The only identifier

int

Username

The user name

varchar

Password

Password

varchar

Phone

Communication number

numbei

Email

Email address

varchar

Sex

Sex

varchar

Robot_type

Model and type of industrial robot

varchar

GroupID

User group unique identifier

int

According to the same structure, the basic data information of industrial robot to be simulated and trained and the user authority information of remote control system

244

D. Zhao and M. F. Qu

can be constructed corresponding database table. During the operation of the system, the links between the database and the system interface, between the database and the software program are formed to ensure that the remote control function can directly call the data information in the database and display it on the human-computer interface in time [7]. At the same time, the real-time remote control information can also be stored in the system database in time.

4 Software Function Design of Simulation Training Remote Control System The simulation training remote hardware system of industrial robot is the hardware foundation of software function realization, and the system database provides sufficient data support for software function realization. In this case, the mobile function and picking function of industrial robots are simulated respectively, and the remote control of the two simulated training processes is realized [8]. In the process of software function realization, deep learning algorithm is applied to realize the remote control function of industrial robot simulation training. The specific software function realization structure is shown in Fig. 3. Human-computer interaction interface Robot control command

Robot control command

Robot dynamics model

Robot dynamics model

Drive system

actuator

Data processing system

Drive system

actuator

Fig. 3. Software function architecture

4.1 Establishing the Kinematic Model of Industrial Robot In order to build the kinematic model of industrial robot, the following assumptions are made as follows. First, the structure of mobile robot is rigid; Secondly, the motion plane of mobile robot is plane;

Simulation Training Remote Control System of Industrial Robot

Rigid structure

Plane of motion

The center of rotation coincides with the center of mass

245

Same resistance, no relative sliding between wheel and ground

Fig. 4. Kinematic model of industrial robot

Third, the rotation center and the mass center of the mobile robot are coincident; Fourth, the resistance of the left and right wheels of the mobile robot is the same. There is no relative sliding between the wheels and the ground. Some factors affecting the speed, such as the quality of the mobile robot, the friction resistance of the ground, and the load-bearing load, are ignored [9]. Then in XY coordinate system, the mobile coordinate of industrial robot is defined as (xR , yR ), the heading angle of mobile direction is expressed as θR , and the initial pose of industrial robot is expressed as (xR , yR , θ ). √ (1) Rmax = 2d 2 The relationship between linear velocity and angular velocity of mobile robot is expressed as follows: ⎧     r r ⎪ v ωl ⎪ 2 2 ⎪ = ⎪ r r ⎪ ⎨ ⎡ω ⎤ ⎡− d d ⎤ωr ⎤ ⎡ |v| · cos θ cos θ 0   x ⎪⎣ ⎦ ⎣ v ⎪ ⎪ y = sin θ 0 ⎦ = ⎣ |v| · sin θ ⎦ ⎪ ⎪ ω ⎩ θ |ω| 0 1 (2)

T Where, [v, ω]T represents the control vector of industrial robot, and x.y.θ represents the state vector of industrial robot simulation training. 4.2 Remote Control Function of Industrial Robot Mobile Simulation Training 4.2.1 Detect Obstacles in Training Environment The collected simulated mobile training environment image is divided into two parts: foreground and background. By modeling the background, the difference between the current frame and background image is used to determine the foreground target area, which is particularly suitable for foreground obstacle detection. The specific detection process is shown in Formula 3.  1, |f (x, y) − f0 (x, y)| ≥ T (3) D(x, y) = 0, |f (x, y) − f0 (x, y)| < T Where, f (x, y) and f0 (x, y) are the current image and background image respectively, D(x, y) is the difference image between the current image and background image, and t is the binary threshold. After background subtraction is used to remove the interference

246

D. Zhao and M. F. Qu

of the ground and the perspective, the general contour of the foreground obstacles can be obtained. However, due to the noise and hollowness of the obtained difference image, the edge of the obstacles is not smooth, so it needs to carry out morphological processing to denoise. After three steps of contour extraction, area filtering and convex hull processing, the accurate obstacle area is finally obtained. 4.2.2 Deep Learning Planning for Obstacle Avoidance of Robot According to the principle of deep learning, decision function, evaluation function, reward function and external environment are the basic components of deep learning system [10]. The decision function of agent is the core of the whole deep reinforcement learning system, which directly determines the action of agent. Decision function π(a|s ) represents a mapping from state to behavior. The decision expression of agent at time t is as follows:

(4) π (a|s ) = P at = a|s t = s Deep learning seeks the best decision by continuously training agents. According to the above theory, the mobile route of industrial robot is planned according to the deep learning framework shown in Fig. 5.

Perceiving the environment gets the state s

Ste p pre dict ion

Decision making

Industrial robot

Reward function

Ma xim um rew ard acti on

To perform a

Action space

The environment

Robot state and surrounding obstacle state

Fig. 5. Deep learning framework of robot movement

According to the learning framework shown in Fig. 4, the straight-line interpolation algorithm in Cartesian coordinate system is used to plan the captured and placed straight-line tracks. The middle track is still planned by quintic polynomial interpolation algorithm, and the workpiece is located in three-dimensional real-time. Using the binocular vision of ABB Robot, the working area images of left and right cameras are collected respectively, and then the workpiece is located by threshold segmentation and Hough transform Finally, the height information is obtained by stereo matching. The coordinates of the workpiece are recorded as A0 (x0 , y0 , z0 ), and the end point of the trajectory, i.e. the placement point, is known and set as C(x2 , y2 , z2 ). Using the background subtraction method, the image area of the obstacle can be determined by subtracting the background image from the working area image of the left camera, and the center point coordinate is recorded as B(x1 , y1 , z1 ). The height of grasping is determined to avoid obstacles. The straight-line trajectory is parallel to z-axis of robot, including two tracks

Simulation Training Remote Control System of Industrial Robot

247

of grasping and placing. A point is inserted every 0.2 s by using the timing interpolation method. The straight-line trajectory moves at a constant speed with a speed of 5 cm/s. all interpolation points are inverse solved. The intersection coordinates of the straight-line trajectory and the middle trajectory are determined and the inverse solution is obtained. Linear interpolation and quintic polynomial are combined to fit the trajectory of each joint. Every joint of the robot is controlled regularly. All the joint angles of the robot obtained above are transformed with time and sent to the robot at the same time every 0.2 s to control the motion of each joint. 4.2.3 Transmission of Mobile Training Control Instructions for Industrial Robots The transmission of control instructions can be divided into two parts. One is to collect mouse coordinates under the supervision and control mode, and use the mouse to set the destination. The mouse coordinates to be transmitted at this time are the control instructions. The other is the acquisition of the key code clicked by the mouse in the direct control mode. The key code to be transmitted is the control instruction. What we need to achieve now is to transfer the control instructions of the server to the robot control server through the wireless network through the middleware technology based on. 4.3 The Remote Control Function of Grab Picking Simulation Training for Industrial Robot According to the same way, first use the deep learning algorithm to detect the target, and the specific principle is shown in Fig. 6.

Deep learning error function The error gradient The current action The environment

Current value network The current state Current status and actions

Copy the parameters every N steps

Target value network

Next state Playback memory unit

sample

Fig. 6. Principle diagram of target detection by deep learning algorithm

The output of the target detection algorithm can be expressed as follows:   P b = pnb , . . . , pib

(5)

Among them, n represents n targets after detection from the image and 0.7 threshold filtering. pib represents the probability of the target with index I. Through the position

248

D. Zhao and M. F. Qu

information of the output box of the object detection results to be grabbed, n depth regions of interest can be obtained from the corresponding depth estimation results. Then control and control industrial robot serialization grab, control flow as shown in Fig. 7.

Target detection

Next loop

The depth of the estimated

Candidate fetching probability calculation

Coordinate transformation

Serialized grab control strategy

Industrial robots perform the pick and grab process

Critical point estimation

Fig. 7. Control flow chart of industrial robot serialization and grabbing

5 System Test 5.1 Choose Simulation Training Industrial Robot In the test experiment, the simulated training industrial robot is aubo-i5, a light manmachine cooperative robot developed by Beijing intelligent. The model of industrial robot is based on the concept of modularization and adopts an open software architecture. Aubo-i5 provides a variety of interfaces based on ROS. For the inverse solution and trajectory planning of manipulator kinematics, we can use not only the integrated moveit package of ROS, but also the interface of manipulator motion provided by Aobo company. Robotiq-140 claw is used as mechanical claw, which is developed by robotiq company of Canada. 5.2 Remote Control System Function Test 5.2.1 Mobile Control Function Test In order to verify the operation effect of the robot obstacle avoidance module, the simulation test is carried out on the simulation software stage provided by ROS. Stage is a relatively simple 2D simulator, which can simulate a single or multiple robots, and add robots by modifying the file at the end of. World. It can load a variety of manually designed maps and add the desired obstacles. The designed test environment is a 10 × 10 area, the actual length of one grid is 1 m, the robot is set as a circle with a radius of 0.5 m, the maximum moving speed of the robot is set as 1 m/s, and the trajectory arc is set as red. The gain coefficient of gravitational field is set to 1, the gain coefficient of repulsive field is set to 6, and the influence range of obstacles is set to 2. Aiming at the problems of collision, target inaccessibility and local minimum point caused by too

Simulation Training Remote Control System of Industrial Robot

249

Table 2. Configuration success interface of industrial robot The serial number

Traditional industrial robot simulation training remote control system

Remote control system of industrial robot simulation training based on deep learning

The experimental results

Collision times of obstacles

The experimental results

Collision times of obstacles

1

Successful

2

Successful

0

2

Successful

2

Successful

1

3

Failure

4

Successful

2

4

Successful

3

Successful

1

5

Successful

1

Successful

0

6

Failure

4

Successful

2

much gravity, the simulation test is carried out. The test results are obtained by comparing with the traditional industrial robot simulation training remote control system, as shown in Table 2. From the experimental results shown in Table 2, it can be seen that compared with the traditional remote control system, under the control of the designed remote control system, the success rate of mobile control of industrial robot has reached 100%, and the number of collisions with environmental obstacles has been reduced by 62.5%. 5.2.2 Picking Control Function Test Figure 8 shows the specific scene of detecting the picking control function of the remote control system.

Fig. 8. Schematic diagram of system picking control function test

The experiment flow is controlled by code on the computer side. Firstly, Kinect collects the color and depth map of the scene, sends the color map into the network to predict the capture center point and the capture direction, then calculates the values of the capture point and the capture direction in the base coordinate system according to

250

D. Zhao and M. F. Qu

Kinect’s conversion relationship, and finally sends these values into the control program of the manipulator, which is executed by the manipulator. Compared with the traditional remote control method, it is found that under the simulation training remote control system of industrial robot based on deep learning, the speed of industrial robot picking up the designated objects is faster and the rate of empty grasping is lower. 5.3 Remote Control System Performance Test Real time is an important performance index of any remote control system. There are many reasons that affect the real-time performance of remote control, such as hardware performance, software performance and network transmission. In the remote control, due to the existence of many factors, the industrial robot may not receive the control command from the client immediately, resulting in the robot can not respond to the control of the client quickly. To some extent, this causes the response delay of mobile robot, which affects the user control experience and the normal operation of the whole remote control system. Because of the relationship between response time and real-time, the shorter the response time is, the higher the real-time control is. The response time of the control command of the remote control system is calculated by recording the time when the control command is sent out and the time when the industrial robot executes the command. The comparative experimental results are obtained by statistical calculation, as shown in Table 3. Table 3. Comparison results of control delay of remote control system The serial number Traditional industrial robot simulation training remote control system

Remote control system of industrial robot simulation training based on deep learning

1

2

3

4

Control command issuing time

11:16:20

11:18:10

11:20:45

11:24:01

Industrial robot response time

11:16:47

11:18:45

11:21:13

11:24:34

Time difference/s

27

35

.38

33

Control command issuing time

11:16:20

11:18:10

11:20:45

11:24:01

Industrial robot response time

11:16:39

11:18:29

11:21:08

11:24:24

Time difference/s

19

19

23

23

From the data in the table, it can be found that the average time delay of traditional remote control system is 33.5 s, while the average time difference of the designed remote control system is 21 s. It can be concluded that the design method can save about 12.5 s on average in the aspect of control delay performance.

Simulation Training Remote Control System of Industrial Robot

251

6 Concluding Remarks The simulation training remote control system of industrial robot based on deep learning is the product of the integration of the development of machinery, computer technology and communication technology. This kind of industrial robot can adapt to the working environment of various adverse conditions, which can not only improve the working efficiency, but also reduce the cost and improve the working quality. So in all walks of life, this kind of networked, integrated and intelligent industrial robot will be more and more widely used. The system has high reliability and secondary development of communication, which can save about 12.5 s running time on average. It can not only monitor the working state of industrial robot in real time, but also improve the running efficiency, so it has high application value. Fund Projects. Key topics of Beijing Polytechnic, Research and design of equipment management system based on RFID (CJGX2016-KY-YZK041).

References 1. Zhou, Z., Huang, G., Gao, J., et al.: Radar emitter identification algorithm based on deep learning. J. Xidian Univ. 44(3), 77–82 (2017) 2. Kamrava, S., Tahmasebi, P., Sahimi, M.: Enhancing images of shale formations by a hybrid stochastic and deep learning algorithm. Neural Netw. Off. J. Int. Neural Netw. Soc. 118, 310 (2019) 3. Keqiang, B., Zhigui, L., Yingtong, W.: An integrated design method coupling structure and control for industrial robot. Sci. Technol. Rev. 36(9), 91–96 (2018) 4. Vuthi, Y., Wangyao, N., Phamvan, C.: Recurrent fuzzy wavelet neural networks based on robust adaptive sliding mode control for industrial robot manipulators. Neural Comput. Appl. 31(18), 1–14 (2018) 5. Rosa, D.G.G., Feiteira, J.F.S., Lopes, A.M., et al.: Analysis and implementation of a force control strategy for drilling operations with an industrial robot. J. Braz. Soc. Mech. Sci. Eng. 39(1), 1–8 (2017) 6. Xiuxing, Y., Li, P.: Direct adaptive robust tracking control for 6 DOF industrial robot with enhanced accuracy. ISA Trans. 72, 178–184 (2017) 7. de Gea Fernández, J., Mronga, D., Günther, M., et al.: Multimodal sensor-based whole-body control for human-robot collaboration in industrial settings. Robot. Auton. Syst. 94, 102–119 (2017) 8. Santos, J., André, C., Santos, T., et al.: Remote control of an omnidirectional mobile robot with time-varying delay and noise attenuation. Mechatronics 52 (2018) 9. Lee, D., Lee, J.: A hybrid joystick with impedance control for a stable remote control of a mobile robot. Int. J. Human. Robot. 16(1) (2019) 10. Santos Lopesdos, M.S., Gomes, I.P., Trindade, R.M., et al.: Web environment for programming and control of a mobile robot in a remote laboratory. IEEE Trans. Learn. Technol. 10(4), 526–531 (2017)

Knowledge Training System of Urban Pest Control Based on Big Data Analysis Xin Zhang(B) and Ming-fei Qu Beijing Polytechnic, College of Mechatronic Engineering, Beijing 100176, China [email protected]

Abstract. Urban insects are a large group of arthropods closely related to urban economic construction and people’s life. If they are not effectively controlled, they will bring harm and loss to urban economic construction and people’s life. In this context, in order to help the relevant urban management departments to better complete the work of urban pest control, the design of urban pest control knowledge training system based on big data analysis is studied. The research includes the overall design and detailed design of the system. The overall system discusses the overall system architecture, system function architecture, system architecture, system topology and other macro or surface theories. The technical points of the whole training information management system are in the detailed design stage, which is divided into system login configuration module, training demand management module, resource management module, teaching management module, performance management module, system management module and other modules. Finally, the system test and performance analysis are carried out for the system. The results show that the training system can support 500 users to use simultaneously and meet the needs of users’ learning and training. Keywords: Big data analysis · Urban pests · Knowledge of killing and prevention · Training system

1 Introduction With the development of the city, the ecological environment of the city, including climate, hydrology, animal and plant communities, has changed. With the increase of houses, warehouses, processing plants and restaurants, more shelters have been provided for cockroaches, beetles, longicorn and other pests. The domestic garbage and waste are not cleaned up in time, and the number of pets has expanded, providing a rich source of food for flies, mosquitoes, fleas and other pests. At the same time, with the global warming and the widespread use of indoor air conditioning, the weather in the city tends to be rainy and humid, which provides rich food and suitable living place for ants, cockroaches and other pests. The possibility of large-scale insect disaster in the city is greatly enhanced, and the cockroaches, known as “rich and valuable insects”, have replaced bedbugs and lice to become the first of the new four pests [1]. Urban © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 252–264, 2020. https://doi.org/10.1007/978-3-030-63955-6_22

Knowledge Training System of Urban Pest Control

253

pests are closely related to people’s production and life, which cause great harm and loss to people’s health, food, architecture, furniture and storage. Therefore, more and more attention has been paid to them, and people’s desire for urban pest management is becoming stronger and stronger. Killing and controlling urban pests is not only effective by individuals, but also requires the collective efforts of relevant urban management departments. However, the reality is that people’s knowledge of urban pest killing and control is relatively weak, especially the use of some insecticides is more vague, once used incorrectly, the consequences will be unimaginable [2]. Therefore, it is not necessary to train the relevant staff on the knowledge of urban pest control. At present, the training of urban pest control knowledge mainly relies on human propaganda and collective activities. However, the training effect of these methods is very small, which not only limits the scope, but also wastes a lot of human and material resources. Therefore, it is of great practical significance to study a training system of urban pest control knowledge, which not only breaks the limitation of training time and space, the content of the training is more standardized and comprehensive, playing the effect of joint participation [3]. However, the traditional urban pest control knowledge training system can not support more users to use together, and can not meet the needs of user learning and training. Aiming at the problems of the traditional system, this paper designs a knowledge training system of urban pest control based on big data analysis. Based on big data analysis, the overall design and detailed design of urban pest killing and control knowledge training system are two levels. The system overall discusses the system architecture design, overall structure design and system function module design. The detailed design of the system discusses the six functional modules of the whole training system. At last, we find a way to test the whole training information management system, and test the concurrent performance of the system. The results show that the training system can support 500 users to use simultaneously and meet the needs of users’ learning and training.

2 Design of Knowledge Training System for Urban Pest Control Urban pests, in a narrow sense, refer to insects that live in urban environment and cause harm or inconvenience to human production, life and health, such as mosquitoes, flies and termites; in a broad sense, urban pests refer to those that live in urban environment and cause danger to human production, life and health Harmful or inconvenient animals, such as insects such as mosquitoes, flies and termites, as well as ticks, mites, scorpions, spiders, centipedes, mice and other mammals and arthropods [4]. The development of the city will inevitably lead to the reduction of the land for agriculture and forestry and the destruction of a large number of natural vegetation, while the land for agriculture and forestry has become a part of the urban division and the reduction of natural vegetation around the city, which in turn affects the non biological factors (soil, climate, etc.) in the city; in addition, the complexity of urban construction makes the urban environment complex and changeable. On the one hand, these changes bring environmental pressure to the survival of urban pests, which makes their living environment severely disturbed; on the other hand, they provide abundant habitat resources for the survival of urban pests, which makes the species of pests more

254

X. Zhang and M. Qu

abundant. The richness of pest community in urban environment is high, and it has significant heterogeneity, which is related to the complex diversity of urban environment. For example, the pests in urban areas are different from those in suburban areas. Due to frequent commercial activities, high concentration of people and objects, strong mobility, pests in urban areas are generally more than those in suburban areas; buildings in urban areas are highly concentrated, and there are less vegetation such as water surface and pond, and pests such as mosquitoes and flies are generally less than those in suburban areas. In the face of the above situation, urban pest control is imminent. 2.1 System Overall Design Considering the research and system requirements of the system, the basic requirements of online training learning in the design process are shown in Table 1 below. Table 1. Overall design requirements of the system Basic requirements Explain Stability

The system should not go down as much as possible, and it can have high reliability on the premise of meeting the concurrent needs of users

Foresight

The training system developed has some advanced nature to ensure that it will be in the lead level in the next two to three years

Applicability

It has universal applicability to all relevant staff while achieving the expected effect of training for relevant staff

Extensibility

The system must be able to upgrade conveniently according to the business development and users’ needs. The architecture design and development platform of the system should have good scalability and portability

2.1.1 System Architecture Design The system architecture design is divided into five layers: infrastructure layer, data resource layer, application support layer, business implementation layer and web client access layer. The specific architecture is shown in Fig. 1 [5]. (1) Infrastructure layer It mainly involves system infrastructure, including hardware server, network facilities, operating system, database system, software, etc., as well as relevant standards and specifications to ensure the stable and efficient operation of the system. (2) Data resource layer All kinds of data files related to system operation, such as external files, business data, logical data, etc., are stored in different database servers. (3) Application support layer Technical support of system application, including workflow engine, form designer and web service.

Knowledge Training System of Urban Pest Control

Web client access layer

Administrator, new employee, instructor, training administrator, auditor

Business impleme ntation layer

Identity verification

Unified login

System login configuration, training demand management, resource management, teaching management, score management, system management

Application support layer

Workflow engine

form designer

Data resource layer

Business data

Documentation

Infrastr ucture layer

255

Hardware

Network facilities

Web Service

External data

Systems software

Database software

Fig. 1. Overall system architecture

(4) Business implementation layer The level of realizing business functions, including system login configuration, training demand management, resource management, teaching management, performance management, system management and other business functions, is the core part of the whole system. (5) Web client access layer The overall architecture of the system, B/S mode to the user browser as the client. B/S is the window of interaction between application system and users. Users can access it through HTTP protocol. 2.1.2 System Function Architecture Design The training system of urban pest killing and control knowledge has six functions: system login configuration, training demand management, resource management, teaching management, performance management and system management. The main functional framework of the urban pest killing and prevention knowledge training system is shown in Fig. 2 [6]. 2.1.3 System Architecture Design Because the structure of B/S belongs to the progress and innovation of architecture, the web-based mode is the response of application service to client’s request service. Through the browser interface, the client can access and handle the business, and the business logic is processed on the web server. At present, the B/S structure is a more popular solution of the system architecture. The web server is located in the business logic layer in the B/S structure. It can process the requests submitted by users, and access

256

X. Zhang and M. Qu System login configuration

training demand management

resource management

Knowledge training system of urban pest control

teaching management

score management

system management

Fig. 2. System functional architecture

the data in the data layer. According to different requirements, it needs to update the database [7]. The design mode of the urban pest killing and control knowledge training system is shown in Fig. 3.

user browser

User request

Server response Page orientation

Control layer

View layer

instantiation Model level

Web container / server

Database operations

data base

Fig. 3. Design mode of urban pest killing and control knowledge training system

Knowledge Training System of Urban Pest Control

257

The overall architecture of the staff training management system includes data layer, business logic layer and presentation layer. Each layer of the architecture is interrelated. The three-layer architecture is independent and interrelated in each function. In the framework of employee training management system, application server and database server need to be configured. The application server processes business logic after responding to the client’s request. During the process, data call is required to the database server. After the application server processes business logic, the server displays the results to the web of client browser. The system that generates the results from the application program, displays the designed interface and unified standard B/S mode, does not need to consider the physical location of the system database, and all the complex work in the system is uniformly handled by the application server [8]. 2.1.4 System Topology Design The training system is based on the B/S mode, and the system architecture adopts the three-tier architecture of database server, application server and browser. The application server is used to deploy the training system program of urban pest killing and prevention knowledge. In this paper, the application server adopts redundant design. When one server fails or has problems, the other can take over completely. If both servers operate normally, the load balance can be realized. The database server runs the database management system of this paper, and the storage device is used for storage System data information. Because the training system is in the Internet environment, and needs to provide 7 × 24 h of operation support, so the network design and deployment needs higher requirements. The important hardware facilities in the network (such as core switch, core server) need to introduce redundancy scheme in the network design, and can be directly replaced in case of failure. The network topology design of online training and learning system is based on eliminating single point fault as much as possible, and the important equipment has redundancy scheme. For example, in the network of this paper, we deploy two cross switches for information exchange and transmission. The application server and database server adopt the form of two computers, which can not only realize standby, but also realize the load balance of resources when the system is under great pressure. For example, the application server adopts redundancy design, two devices are installed with application programs, one as standby. When one application server has problems, it can be directly switched to another standby server for use, so as to improve the availability and reliability of the system. RAC load balancing is adopted for database server. Under normal circumstances, any server can process data. Once one server has sudden accident or failure, the other server can operate independently without affecting the normal operation of the system. 2.2 System Software Design From the functional structure of the system in the previous chapter, we can clearly see that the system mainly includes six functional modules. In this chapter, each functional module will be designed and described in detail.

258

X. Zhang and M. Qu

(1) System login configuration module design For this module, we need to define the relevant initial data, so that the whole configuration of the system can be initialized. This work is generally carried out by the system administrator. One of its core functions is to make a preliminary plan for permissions, plan roles, set related passwords, and set parameters for related system column menus [9]. (2) Training demand management module design Training demand management function business mainly refers to the definition and acquisition of training demand, as well as the filling and review of demand. Training demand management includes the functions of demand survey definition, Department demand filling, Department demand review, employee demand filling, employee demand review, training demand query, etc., mainly covering the operations of adding, deleting, modifying, checking, filling, submitting, reviewing, printing, etc. for various functional businesses. The functional architecture of the training demand management module is shown in Fig. 4. Requirements survey definition

Department requirement filling

Training demand managemen t module

Department requirements review

Employees are required to fill in

Review of employee requirements

Training demand query

Fig. 4. Functional architecture of training demand management module

(3) Design of resource management module The main function of this function module is to prepare for the next training and learning around the relevant resources of training and learning, including the upload and update of learning materials such as courseware materials, documents, courseware videos, etc. Before the training, the training master data information must be entered and managed. The existence of training resources is an important information resource to ensure online training and learning. It is the basis to ensure that the system training can meet the expectations of the trainees. According to the description in the system requirements specification, we design the training resource

Knowledge Training System of Urban Pest Control

259

management module as the following parts. The functional structure design is as follows: Fig. 5 shows [10].

Category design

Courseware management

Library management Resource managemen t module

Lecturer management

data management

job management

Fig. 5. Functional architecture of resource management module

(4) Teaching management module design Teaching management includes six sub modules: course management, online homework, online learning, online examination, score query and teaching evaluation, and its functional framework is shown in Fig. 6. (5) Test management module design In this module, the main content is to test the learning effect of users, and use the results to carry out subsequent evaluation and course improvement. According to the test paper corresponding to the user training course, when the user completes the training, submit the test paper automatically matched with the background setting to take part in the exam and answer questions. According to the set answer time in the answer, the user can submit the test paper manually within the corresponding time, or the system will automatically submit the test paper when the answer time is too long. And according to the test results to generate test scores. The functional architecture of the exam management module is shown in Fig. 7. (6) System management module design System management business function refers to the management and maintenance function of system basic information such as system parameters, training categories, training reports, etc., mainly including the statistical analysis of training situation of relevant staff, as well as the management of query sets defined and assigned permissions in the engine. The functional architecture of system management is shown in Fig. 8.

260

X. Zhang and M. Qu

course management

Online homework

Online learning Teaching manageme nt module Online examination

Score query

Teaching evaluation

Fig. 6. Functional framework of teaching management module

Knowledge point management

Item management

Item bank management

Test managemen t module

Examination paper management

Examination arrangement and monitoring

Manual review

Score management

Fig. 7. Functional architecture of examination management module

Knowledge Training System of Urban Pest Control

261

Parameter setting

Training category management System managemen t module

Training report management

Common query analysis

Fig. 8. Functional architecture of system management module

3 System Implementation and Test After analyzing the requirements of the system and designing the system, the next step is to realize the system. This chapter introduces the implementation of the system. 3.1 System Development Environment The system configuration of hardware and software environment of employee training management system is as follows: (1) System hardware environment The hardware environment of the system adopts Intel Core i5 CPU, 500 g hard disk and 4G memory. (2) System software environment Using Windows 7 operating system as the system design and client test platform, this paper uses MyEclipse 5.1 development kit, Tomcat 6.0 server, SQL Server 2005 database, JDK 1.5 3.2 Test Tools In the whole test process of the urban pest killing and prevention knowledge training system based on big data analysis, the function test is mainly manual test, and the performance test is mainly test tools. The test tools we used include: (1) LoadRunner: a commercial stress testing tool. It simulates virtual users to perform concurrent operation of the system, and records the operation of all aspects of the system. Testers can also monitor the system load in real time. (2) HP QuickTest professional: auxiliary system function test tool. With QTP tool, testers can record scripts, edit scripts, and run test scripts. (3) Bugzilla: it is an open-source free error or defect tracking system, which can help testers manage the whole process of software development, including defect submission, defect repair, defect closure, etc.

262

X. Zhang and M. Qu

3.3 Performance Testing One of the biggest problems in the design of this system is how to meet the needs of a large number of users logging in at the same time to complete online learning and training. Therefore, the performance test of this system is the concurrent performance test of the system. First of all, through the establishment of a large number of test users, in the LoadRunner tool, by setting to load a user every 3 s, until all users are loaded, all users concurrently perform system business operations, and the performance test report of urban pest killing and prevention knowledge training system based on big data analysis is shown in Table 2. Table 2. Training system performance test report Number of concurrent users

System average response time (s)

Application service CPU usage (%)

CPU usage of database server (%)

Number of business errors

100

0.006

2.1

1.3

0

200

0.014

2.2

1.3

0

300

0.019

2.3

1.6

0

400

0.04

4.6

2.1

0

500

0.07

5.3

2.7

0

600

0.32

6.3

3.6

0

It can be seen from Table 2 that the response of concurrent users performing business operations between 100 and 600 is 0.006–0.4 s, the CPU utilization rate of application server is 2%–7%, and that of database server is 1%–4%. When the number of concurrent users in the system reaches 500 peak, the CPU utilization rate of application server and database is 0.006–0.4 s The utilization rate is in the normal range. At the same time, we also conduct real-time monitoring on the important performance indicators of the database and application server, such as disk read-write and memory usage, as shown in Figs. 9 and 10.

Knowledge Training System of Urban Pest Control

263

Fig. 9. Application server performance during stress test

Fig. 10. Database server performance during stress test

According to Figs. 9 and 10, the resource utilization rate of both is within a reasonable range. Through the test, it can be concluded that the training system can support 500 users to use simultaneously.

4 Concluding Remarks In conclusion, the rise of urban entomology in China has become a necessity due to the acceleration of urbanization. With the rise of Urban Entomology and the further modernization of the city, the problem of urban pest management will be put on the agenda. Therefore, this paper studies the knowledge training system of urban pest control based on big data analysis. The system has been tested to meet the needs of a large number

264

X. Zhang and M. Qu

of users, but there are many deficiencies in the design and development process of the system. In the future research, in order to make the relevant staff training system play its auxiliary role in urban pest control, we need to optimize the system designed in this paper in the following aspects, such as how to establish a reasonable curriculum system, how to arrange the relationship between training programs, how to realize the quantitative assessment of online test results, how to avoid the score test One sidedness in evaluation and singleness in personnel training. Fund Projects. Key topics of Beijing Polytechnic, the application of UAV spraying in the city pest control service (2019H033-KQ).

References 1. Courtierorgogozo, V., Morizot, B., Boëte, C.: Using CRISPR-based gene drive for agriculture pest control. 18(9), 1481 (2017) 2. Amin, P.W.: A modified ancient prescription for crop yield improvement and cotton pest control. 6(2), 163–166 (2018) 3. Grzywacz, D., Moore, S.: Chapter 7 – Production, formulation, and bioassay of baculoviruses for pest control, 109–124 (2017) 4. Gagic, V., Paull, C., Schellhorn, N.A.: Ecosystem service of biological pest control in Australia: the role of non-crop habitats within landscapes: native vegetation and biocontrol. Austral Entomol. 57(2), 194–206 (2018) 5. Xie, X., Shi, L., Zhang, Z., et al.: Multiservice training simulation system for multilevel power grid dispatching and operation part one system design and implementation. Autom. Electr. Power Syst. 41(13), 100–105 (2017) 6. Zhang, B., Wang, L.: Real-time simulation training system for substation based on FPGA. Power Syst. Protect. Control 45(6), 55–61 (2017) 7. Hou, Y., Lin, Y., Shi, J., et al.: Effectiveness of the thoracic pedicle screw placement using the virtual surgical training system: a cadaver study. Oper. Neurosurg. 15(6), 6 (2018) 8. Chen, Z., Wang, L.: Reconstruction of a clinician training system in China—a successful “5 + 3” model from Shanghai. 32(3), 264–269 (2017) 9. Yang, W., Wang, H., Xu, Y., et al.: Effect of balance evaluation and training system on early trunk control ability of patients with cerebral apoplexy. Rehabil. Med. 27(2), 17 (2017) 10. Oshita, M., Inao, T., Ineno, S., et al.: Development and evaluation of a self-training system for tennis shots with motion feature assessment and visualization. Vis. Comput. 35(3), 1–13 (2019)

Research on Educational Data Mining Based on Big Data Xuping He, Wensheng Tang(B) , Jia Liu, Bo Yang, and Shengchun Wang College of Information Science and Engineering, Hunan Normal University, Changsha 410081, Hunan, People’s Republic of China [email protected], {tangws,yangbo}@hunnu.edu.cn, [email protected], [email protected]

Abstract. Educational data mining (EDM) is a cross-disciplinary technology involving computer science, education, statistics, etc. It analyzes and mines education-related data to discover and solve various types of education problems. To make them better understand students and their learning environment and improve the teaching effect of teachers. Under the background of big data, EDM research will usher in a new development space. This paper first analyzes the latest research status of EDM at home and abroad, and then focuses on the progress of EDM in the context of big data in recent years. It summarizes the characteristics, shortcomings and development trends of EDM in the context of big data. Finally, it discusses the opportunities and challenges faced by EDM in the era of big data. Keywords: Big data · Education data mining · Education big data

1 Introduction Education data mining is a comprehensive use of mathematical statistics, machine learning and data mining techniques and methods to process and analyze education big data. Through data modeling, we can find the correlation between learners’ learning results and learning content, learning resources, teaching behavior and other factors, and predict learners’ future learning trend. In recent years, with the development of a large number of educational applications such as online education platforms, social software, mobile phones, etc., which This work was supported in part by the Hunan Province’s Strategic and Emerging Industrial Projects under Grant 2018GK4035, in part by the Hunan Province’s Changsha Zhuzhou Xiangtan National Independent Innovation Demonstration Zone projects under Grant 2017XK2058, in part by the National Natural Science Foundation of China under Grant 61602171, in part by the Scientific Research Fund of Hunan Provincial Education Department under Grant 17C0960 and 18B037, and in part by the Key Research and Development Program of Hunan Province under Grant No. 2019SK2161. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 265–278, 2020. https://doi.org/10.1007/978-3-030-63955-6_23

266

X. He et al.

provide a large number of applications and data for the research of education data mining, the important report of “promoting education and learning through education data mining and learning analysis” issued by the U.S. government in 2012 [2] has triggered an upsurge of application research in the field of big data education. A large number of researches on education data mining and learning analysis, data-driven educational decision-making, large-scale personalized education, adaptive learning systems, and prediction-based teaching interventions have been carried out [2]. Education data mining in the context of big data has become a new research hotspot. Since 2012, EDM research has entered a big data era of “Data driving schools, analysis reforming education” [1]. Big data provides effective improvement means for the lack of equity, accuracy, personalization, innovation and other aspects of the field of education, and has broad development space [2]. At present, there are three active research directions in this field: the first is the study of students dropping out based on the analysis of learning behavior data; the second is the application study of personalized learning services based on recommendation and adaptation algorithm; the third is analysis and prediction of learning behavior based on massive education data mining. This paper focuses on the development of EDM in the context of big data in recent years, summarizes the characteristics, shortcomings and development trend of EDM in the context of big data, and finally discusses the opportunities and challenges faced by EDM in the era of big data.

2 Research Overview 2.1 New Features Education big data mining is a new sub direction of traditional education data mining. The traditional education data mining, because of the less sources of education data at that time, generally came from questionnaires and information management software. The mining method is relatively simple and plays a very limited role in promoting the development of education. With the advent of the information age and the development of online teaching platforms, education data sources have become very extensive: from the perspective of students, it includes life information, learning information and online second classroom information; from the perspective of teachers, it includes teaching tasks, courseware and other teaching information like paper works, scientific research data and other scientific research information; from the perspective of managers, it includes the school’s resources production information, teacher information, enrollment and employment information, etc. At the same time, with the rise of new technologies such as mobile Internet and Internet of things, more and more information is generated by teachers and students and collected automatically by devices [2]. Education has also entered a big data era. Unlike traditional education data, education big data accurately covers all records related to education. Due to the large amount of data, diverse types, strong continuity, and low value density of educational big data [3], the traditional mining algorithms lag behind the big data analysis technology in algorithm efficiency, analysis accuracy, processing heterogeneous unbalanced data and so on, which can no longer meet the mining needs. In addition, the mining method has also changed from a single statistical analysis to the

Research on Educational Data Mining Based on Big Data

267

use of visualization Big data mining, such as clustering, regression, text mining and deep learning neural network, and has developed into education big data mining. Education big data mining can discover the essence of education problems more accurately and effectively, and then promote the development of intelligent education. 2.2 Literature Analysis In order to understand the current research situation of education data mining, this paper searches “educational data mining” as the theme in the web of science database for the period of 2005–2019. Firstly, it makes a comparative analysis on the number of documents, and uses the function of classifying and displaying the retrieval results according to the year in the database to classify the retrieved documents, and obtains the large number of documents at home and abroad according to the number of research results each year when technology is applied to education and teaching. Since 2005, there have been 1421 papers published by EDM, as shown in Fig. 1. Before 2008, the number of papers published by EDM research was relatively small. Since 2008, the number of papers published has increased gradually, especially in recent years. The first International Education Data Mining Conference held in Montreal, Canada in 2008 attracted the attention of researchers. In 2012, the United States Department of Education issued a blue book “Promoting Teaching and Learning Through Education Data Mining and Learning Analysis” [4], marking that EDM has been widely concerned. Since 2015, with the advent of the era of big data, researchers have applied the new technology of EDM to online learning platforms such as MOOC and intelligent learning network, and education big data mining has developed rapidly and becomes a research hotspot. 300

244

250

209

200 150 100 50

261

115 41 22 22 22 25 30

145

60 55 76 71

0

Fig. 1. Publications from 2005 to 2019

2.3 Research Status Research in the United States and other countries started earlier. For example, as early as 2009, some scholars pointed out in their paper [5] that big data will bring about

268

X. He et al.

changes in biological research and teaching. In 2012, the U.S. government issued the important report “Promoting Teaching and Learning through Education Data Mining and Learning Analysis” [4], which triggered the upsurge of application research in the field of big data education. Many aspects emerged, such as education data mining and learning analysis, data-driven education decision-making, large-scale personalized education, adaptive learning system, and prediction-based teaching intervention research. It starts a preliminary attempt to transform from “discovery data” to “mining data”, and focuses on the prediction and decision-making functions of education data mining technology. Research in China started late, and there is a large gap in research breadth and depth compared with foreign countries. In the past 10 years, domestic research on EDM has made some progress [1, 8–13]. With the rise of the era of big data, education big data as a subset of big data has also begun to attract the attention of experts in education [8]. Xu Peng et al. [1] interpreted the 2012 report, analyzing education for change. In the era of big data, Chen Chi et al. [9] introduced big data technologies such as EDM and LA, and designed a big data model oriented to the field of online education, providing ideas for the study of big data in the field of online education. EDM has attracted unprecedented attention in 2015. Such as Zhou Qing et al. [10] mainly introduced the research results of EDM from different education environments. [11] introduced the characteristics and development process of education big data, and finally aimed at the current problems and challenges in the development of educational big data in China present six policy recommendations. Yanmei Chai et al. [12] collected relevant literature from the Web of Science database from 2008 to March 2017 for statistical and visual analysis, and personalized learning services introduced related research results. Yu Fang et al. [13] introduced the current state of EDM research in the past 10 years and designed based on the “waterfall model” in the field of software engineering. Based on the “user-centered” EDM application research framework. At present, the research on Education Data Mining in foreign countries is in the stage of in-depth development. How to realize the discovery and prediction of teaching intelligence is the current research topic. The researchers in the field of Education Data Mining mainly include Romero of Cordoba University in Spain, Ryan Baker of Worcester Institute of Technology in the United States [5], Kalian ace of Sydney University in Australia [6], and the researches of these three researchers are the most representative. Spanish scholars Romero et al. [7], as the authority who noticed the role of education data earlier, first defined the concept of Education Data Mining (EDM). At present, the research work of scholars mainly focuses on: (1) personalized learning service application research based on recommendation and adaptation algorithm; (2) learning behavior analysis and prediction based on massive education data mining. With the rapid development of information technologies such as artificial intelligence, the realization of personalized teaching and behavior analysis and prediction that respect individual learning differences has become a new requirement for dynamic adjustment of teaching strategies in the era of big data.

3 Recent Research Progress Under the background of education big data, the teaching process will produce massive data such as log, student behavior and teacher behavior every day. Big data technology

Research on Educational Data Mining Based on Big Data

269

provides new solutions, standards and tools in terms of storage, processing and knowledge discovery, which can help the education field solve many technical problems in dealing with massive data. From the existing literature, the current research focuses on the following aspects: (1) personalized learning service under education big data; (2) the study of learning behavior mining under big data; (3) the study of students who drop out of school under big data; 3.1 Research on Personalized Learning Services The goal of personalized learning recommendation is to correctly understand the individual differences and provide learning guidance in accordance with learners’ learning habits, knowledge mastery, resource preference, learning objectives, log recording data, etc., so that learners can better understand the learning process, promote the utilization of learning resources and improve learning efficiency and individual personality development. Education big data mining provides effective technical support for personalized learning. It uses big data technology to process and analyze massive education data, and then finds some associations and rules existing in education, selects appropriate teaching methods and contents for students’ individual characteristics, recommends corresponding learning contents and learning paths, and realizes “individualized teaching, personalized development”. Traditional personalized learning service recommendation algorithms mainly include content-based recommendation algorithms, collaborative filtering recommendation algorithms and hybrid recommendation algorithms. Among them, collaborative filtering recommendation algorithms are one of the most widely used recommendation algorithms at present. The basic idea of collaborative filtering is to filter by “neighbor set”, which considers that there are similarities between users who choose the same project. In the personalized online learning system, the main research recommends the test questions to the students based on the user’s recommendation. First, it calculates the difficulty of the test questions. Secondly, in order to find the target user neighbor set, similarity between users is calculated. Then, the recommendation results are generated, and finally the recommendation quality is evaluated [20]. The two bottlenecks of collaborative filtering in recommendation process are “data sparsity” and “cold start”. Sparse data refers to the data in which most values are missing or zero. Sparse data is not useless data, but incomplete information. Much useful information can be mined by appropriate means. In view of the shortcomings of the algorithm, most of the strategies are to optimize the algorithm, but cannot fundamentally overcome the shortcomings of the algorithm. The traditional recommendation algorithm can solve the personalized learning recommendation problem when the data volume is small, but with the advent of information and big data era, in order to solve the recommendation problem under the background of large data volume, researchers propose intelligent recommendation based on deep neural network, which is also the hotspot of personalized learning research under the background of big data. The main process of Intelligent Recommendation Based on deep neural network is to continuously learn the characteristics of massive data and construct a model. Neural network can automatically correct the weight bias and other parameters through continuous irrigation data to fit better learning effect and ultimately improve the accuracy

270

X. He et al.

of recommendation. In terms of efficiency, the intelligent recommendation based on the deep neural network uses the trained model to process data, which greatly reduces the computation compared with the traditional algorithm. The traditional unified algorithm needs to modify the framework code to adjust the model in the process of solution, which makes the cost of improvement huge. By adjusting the parameters of the neural network, deep learning can change the model to produce results. This model is more flexible, and it is more suitable to use the principle of deep learning to solve the problem under the background of big data. For example, Zhang Yongfu [21] and others put forward a personalized learning recommendation system based on LSTM. Long short-term memory (LSTM) recurrent neural network is a kind of recurrent neural network. LSTM has better output in sequence prediction and capturing the evolution of user taste. Compared with traditional collaborative filtering recommendation algorithms, LSTM recommendation is better in recommendation performance than traditional collaborative filtering algorithms. Yang Heng et al. [22] proposed a personalized recommendation method based on deep belief networks in the MOOC environment. By deeply mining the demographic characteristics of learners and the attribute characteristics of curriculum resources and combining the characteristics of learners’ learning behavior, a DBN based model of learners’ interest was constructed. Based on the interest model of learners, a personalized recommendation model of learners is constructed by using the method of DBN classification, which integrates the demographic characteristics of learners and the attribute characteristics of curriculum resources, and processes them into the feature vector of learners. When training the DBN classification model, the feature vector of learners and the feature vector of learners’ behaviors jointly affect the updating of model parameters. This method effectively solves the common problems of cold start and data sparsity of traditional collaborative filtering-based recommendation methods. 3.2 Research on Learning Behavior Mining In the era of big data, the analysis of learning behavior can deeply understand the learning habits and learning characteristics of learners. According to the characteristics of students’ learning behavior, teachers can make teaching plans or divide students into learning groups with complementary learning styles to improve learning efficiency. The current research is mainly carried out in the following two aspects: (1) the research on the potential behavior patterns of students based on the optimized association algorithm; (2) the research on the learning behavior mining based on the social network analysis algorithm. Based on Association Rules Mining Algorithm Since the education informatization was put forward, the data in the field of education has been growing, and mining technology for education data has attracted more and more attention and has become a new hotspot of data mining research. Most of the related management systems in colleges and universities only stay in the simple operation stage of the original data storage, but the relationship and influence between the courses cannot be determined. Therefore, it is very necessary to use association rule analysis to go deep into the data surface, carry out correlation mining analysis, and find out the

Research on Educational Data Mining Based on Big Data

271

correlation between students’ data. Association rule analysis is one of the most active research methods in data mining. Its purpose is to find out the association relationship among various items in a data set. Through association rule mining, we can get useful information with potential value hidden in massive student data, and discover the behavior patterns and association relationships of students. The most famous association rule algorithm is the Apriori algorithm proposed by Agarwal. Most of the mining algorithms are based on the Apriorist algorithm. However, with the advent of the era of big data, the Apriorist algorithm is facing challenges in both time efficiency and space scalability. Therefore, researchers proposed an improved algorithm for massive data processing to mine its behavior patterns more efficiently. Lu Xinyuan et al. [23] proposed a frequent and efficient association rule mining algorithm based on domain knowledge for mining the association relationship between the academic performance of middle school students in educational administration data. At present, most association rule mining algorithms mainly focus on frequent itemset mining, but only mining frequent itemset cannot meet people’s requirements for efficient use of results, and in the practical application field, they do not consider the rich domain knowledge closely related to the data itself, resulting in a large number of redundant rules. These rules have been well known by the industry, so that the result of the fruit is not very interesting. Based on the sequential pattern mining algorithm, the Fui-DK algorithm generates frequent candidate sets. Based on the support and confidence of the classical association rule algorithm, the two parameters of utility degree and interest degree are added to get the efficient set of interesting terms. Then, the support, confidence degree and utility degree of the Association rules that meet the conditions are sequenced and output. Finally, the efficient set of interesting terms obtains Interesting association rule results. The experimental results show that the algorithm can reduce the computing time, and the elimination rate of the known association rules in the field can reach 43%, which can help colleges and universities to carry out time-saving and effective education data mining. Luna [24] et al. proposed an evolutionary algorithm to discover association rules of rare categories in the learning management system, which is used to mine the association of students’ learning behaviors on the Moodle platform, and compared the algorithm with other five association rule algorithms. Geigle et al. [25] added a layer of HMM based on hidden Markov model (HMM) to form TL-HMM unsupervised learning of many student behavior observation sequences to discover potential student behavior patterns. Based on Social Network Analysis Algorithms Rabbany et al. [26] use social network analysis algorithms to evaluate the participation of students in the forum in the course management system, such as tracking the subjects replied by students, the number of posts published, etc., so that teachers can quickly understand the hot topics discussed by students. Mohammed Saqr [27] et al. proposed that visual analysis of social network analysis algorithms and quantitative network analysis (concentration measurement) should be used together to analyze the position and role of students in collaborative knowledge sharing, to monitor online collaborative learning, find gaps and traps in application, guide the potential of informed intervention, and design relevant data-driven stem from the information obtained through monitoring

272

X. He et al.

Pre measures and using experimental, observational, repeated measurement designs to evaluate its effectiveness to promote teaching and learning. 3.3 Research on Student Drop-Out Dropout is not only considered as a serious educational problem, but also a serious social problem. In addition to the high risk of unemployment or underemployment, dropout students are more likely to suffer from mental health problems, such as depression, gang participation or other criminal activities. In the era of big data, there have been new achievements and new methods in the studies of dropout. By using the technology of education big data mining, students with high dropout risk can be identified efficiently and accurately, and factors related to the risk of dropout faced by students can be analyzed. The current research is mainly divided into two categories: one is online education dropout research; the other is traditional classroom dropout research 。In recent years, the main literature of student dropout research is shown in Table 1. Research on Online Teaching Platform With the popularization of Internet education platforms, many online education platforms such as MOOC and intelligent learning networks continue to emerge, and many excellent teachers in colleges and universities in China have also opened excellent MOOC courses, which allow students in other colleges and universities to enter the classroom. However, according to the research, the high dropout rate of online students both at home and abroad has become more and more prominent, and the high dropout rate in online education platforms has also attracted the attention of many researchers. Online dropout prediction research predicts whether students will insist on learning or drop out in the next week by analyzing the data of students’ learning activity logs on the platform. Its research significance lies in its use as an intervention by creators of online courses and education researchers, so that researchers can gain insights into the reasons why students drop out, and predict various signals of dropout for online education platforms to create customized intervention strategies for learners. In the early stage, the research on the reasons for MOOC students’ dropping out was based on statistical analysis. Generally, the reasons for dropping out were analyzed by filling in the after-school questionnaire survey and artificial statistical analysis data. The personalized analysis and intervention on the reasons for students’ dropping out could not be carried out in depth, which made the accuracy of prediction results poor. With the increasingly prominent advantages of machine learning methods in big data processing, the application of machine learning algorithms in MOOC dropout prediction tasks has been promoted in the context of education big data. The general process used in the classic algorithm of dropout prediction model can be divided into three steps: (1) Data Preprocessing; feature selection of dropout data, screening out significant features and discarding non-significant features; (2) Dropout model training and tuning; selecting model according to the actual situation of dropout data and specific problems to be solved, such as sample number, feature dimension, and comprehensive consideration of data characteristics. In the optimization problem, we use cross validation to observe the loss curve, test result curve and other analysis reasons. The adjustment parameters are: optimizer, learning rate, batch size, etc.; (3S)

Research on Educational Data Mining Based on Big Data

273

Model effect evaluation; prediction and evaluation with test set, comparative study on accuracy rate, recall rate, F-score and other evaluation indicators. The flow chart is shown in Fig. 2. Data Preprocessing

Dropout model training and tuning

Model effect evaluation Fig. 2. Dropout prediction flow chart

In recent years, research of MOOC dropout prediction mainly focuses on two shortcomings of classical machine learning algorithms. The first one is the improvement of data preprocessing steps, mainly including the processing of feature rule making and the processing of extremely unbalanced data. The second is the improvement of the prediction model, including model fusion to make up for the shortcomings of the model and the new model. In order to improve the data preprocessing, the feature rules are formulated based on classic machine learning methods. This part of the work requires a lot of manual operations and carrying out a variety of complex feature extraction operations manually. When there are many complex linguistic phenomena in the text, the process of feature rule making becomes very difficult. Sun Xia et al. [14] proposed a use of convolutional neural networks to automatically extract useful features from student behavior data in order to solve the problem of manual feature extraction in the past. CNN learned the changes of each category from many data to realize the stable classification of feature changes in the same category. Wang Xiyu et al. [16], based on the data of eight MOOC courses with the largest number of students selected from the dream course platform of National University of science and technology of national defense, extracted three dimensions from the course factors, learners’ own factors and other personnel factors. A total of more than 40 learning data were used to study the prediction of dropouts, and the most helpful behavior data of each course was analyzed. In the extremely unbalanced data processing, because the dropout students belong to a few categories, the data set is unbalanced. Usually, most students continue to study, and only a few students drop out. In this case, the accuracy may be misleading, because most of the default classifiers will get high accuracy, while a few are easy to be ignored. Therefore, in the process of data processing, the processing of data imbalance is particularly important. It is necessary to design a specific algorithm that can focus on a few categories, which is conducive to the improvement of prediction accuracy. In the case of education data, the research activity is focused on resampling algorithms and cost sensitive algorithms. Data resampling modifies the training data set by adding instances belonging to a few classes to generate a more balanced distribution of data classes. Smote is a resampling method

274

X. He et al.

that has been shown to improve the classification of imbalances, especially when used in combination with C45 and SVM [19]. On the other hand, the cost sensitive learning algorithms consider the classification errors in other categories. Cost sensitive learning allocates a higher misclassification cost to a few classes of samples, and a smaller misclassification cost to most classes of samples. In this way, cost sensitive learning improves the importance of samples of a few categories artificially in the process of training the learner, thereby reducing the preference of the classifier for most classes. Chaplet [15] et al. put forward a cost sensitive learning algorithm for data imbalance to punish the given class for error classification and adjust the weight of a few classes of data, so that the algorithm has better accuracy and better false negative rate. This method is superior to the previous algorithm in the kappa value of Cohen. In the improvement of the prediction model, the traditional machine learning method assumes that the probability of students dropping out of school in different time steps is independent, which is inconsistent with the actual situation in the scene, because the state of students at a certain time will be affected by the state of the previous time, making the accuracy of the prediction result poor; sun Xia et al. [14] can’t correlate the study of the state of learning in different time periods. Research on Traditional Classrooms In traditional education, teachers and school administrators have spent a lot of time and energy to reduce dropout, but until now, it still exists in schools. How to efficiently and accurately identify at-risk students in many students, put forward personalized intervention programs in time, and reduce dropout rate is a major problem faced by schools. With the in-depth application of big data mining technology in the field of education, researchers use massive learning behavior data for data mining analysis to identify students with a particularly high risk of dropping out. It can effectively and accurately identify students with high dropout risk, and analyze which factors are related to students’ DROPOUT risk. Teachers can choose targeted methods to intervene, which can greatly improve teachers’ work efficiency and personalized care for students. In recent years, the research on the traditional dropout field mainly includes the algorithm improvement for the mining of interpretable classification rules and the model evaluation algorithm for the dropout field. Xing et al. [17] took k-nearest neighbor, support vector machine and decision tree as the benchmark algorithm, and then proposed a drop out prediction model based on deep learning compared with the benchmark algorithm, and got higher drop out prediction accuracy. Márues et al. [18] proposed a new icrm2 algorithm based on the mining of interpretable classification rules, in order to obtain more accurate and shorter classification rules than other existing algorithms, to obtain better performance, with the classification accuracy as high as 91%. Lakkaraju [19] et al. put forward an evaluation algorithm for dropout prediction, using evaluation algorithm to evaluate the results of SVM, decision tree, random forest, logistic regression and other classification models. The evaluation algorithm has a strict qualitative and quantitative comparison for the classification model, changing the situation that only general evaluation indicators can be used, and the evaluation of dropout field algorithm is very important great significance.

Research on Educational Data Mining Based on Big Data

275

Table 1. Summary of student dropout studies under big data

Online teaching platform

Traditional classrooms

No.

The main method

Results

Literature [14]

Combining CNN and LSTM sensitive learning algorithms

F1 value is 0.948 higher than other algorithms

Literature [15]

Emotion Analysis Scoring Kappa in Cohen beats Algorithm the previous algorithm Improved neural networks with a value of 0.432

Literature [16]

Multi-linear regression model BP neural network model

Literature [17]

Comparison between Higher accuracy than the drop-out prediction model baseline algorithm sour and benchmark algorithm based on deep learning

Literature [18]

An ICRM2 algorithm is proposed

Up to 91% of classification accuracy

Literature [19]

An evaluation algorithm specifically for drop-out prediction

Changed the situation where only common assessment indicators can be used

Forecast accuracy of more than 90%

Model Experiment and Analysis Experimental Settings (1) Problem setting: predict the dropout behavior of Xuetang Online, one of the largest MOOC platforms in China. Through the interpretation of the data set, based on the user’s previous behavior, he predicts whether he will skip class in the next 10 days. The defining criterion is that a registration number has no log record within 10 days after a certain point in time. (2) Data set: The data set used in the experiment is a publicly available data set provided by KDD Cup2015. This dataset contains 120,542 registered activity logs from 79,186 students across 39 courses, and each course can take up to 5 weeks. There are two main sources of learner behavior record information: browsers and servers, including seven events: objects visited, discussions, navigation courses, page closures, attempts to solve problems, watching videos, and browsing the wiki. Each student’s course behavior includes watching videos, trying to solve problems, participating in the course. (3) Evaluation indicators: Accuracy, recall and F1 values commonly used in dropouts are used as evaluation criteria. TP indicates the number of dropouts correctly predicted, FP indicates that the average student was incorrectly predicted as the number of dropouts, and FN indicates the number of students that were predicted to be actually dropouts. Precision, recall and F1 values are calculated by formulas (1), (2)

276

X. He et al.

and (3). precision = recall = F1 =

TP TP + FP

(1)

TP TP + FN

(2)

2 × precision × recall precision + recall

(3)

(4) Experimental model: A comparison analysis between traditional classification methods (LR (logistic_regression), DT (decision_tree), SVM (Support Vector Machine)) and deep learning methods (GBDT (GradientBoostingDecisionTree), MLP (Multilayer Perceptron)). Experimental Results and Analysis Table 2. Comparison of the five dropout prediction models LR

DT

SVM GBDT MLP

Accuracy 0.821 0.798 0.661 0.813

0.873

Recall

0.833 0.815 0.712 0.843

0.892

F1

0.827 0.806 0.686 0.828

0.882

Among these prediction models, the performance of SVM is slightly worse, because the amount of data used in the experiments in this paper is large, SVM is not suitable for large sample data mining, and it is better for small sample SVM classification. When the amount of data is large, MLP performs best as a method of neural networks, which shows that neural networks have superior performance in educational big data mining and are more in line with the current needs of data mining in the context of big data (Table 2). The overall prediction accuracy in the experiment is low. After analysis, the existing research and methods have the following deficiencies: (1) the use of big data cannot be used to improve the prediction accuracy, and the use of fine-grained features of student behavior records is insufficient (2) neural network methods The prediction result can explain too poor (3) Cannot solve the common cold start problem in practical application scenarios.

4 Summary and Outlook After decades of development, EDM has attracted more and more researchers’ attention, especially in recent years, online education platform, social software, mobile phones and so on provide many applications and data for the research of education data mining. In

Research on Educational Data Mining Based on Big Data

277

recent years, EDM research has made great progress, but there are still some deficiencies, mainly reflected in four aspects: first, there is little research on data preprocessing technology in the data preparation stage, only a few scholars put forward some improvement methods. Second, in the process of building the model, deep learning technology cannot make full use of big data to improve the prediction accuracy, and cannot make full use of the fine-grained characteristics of student behavior records and the poor interpretability of the model. Third, in the new teaching environment, there is less research on the mining of teachers’ educational process, which is only limited to the mining and prediction of students’ learning behavior. However, teachers also play an important role in the process of education. Fourth, there is a lack of open data set in the aspect of information sharing, in which the disclosure of students’ information may have problems such as privacy and security, and the lack of data sharing It hinders the research of education data mining, which leads to each research aiming at the specific teaching scenes collected by itself, and fails to conduct research in more and wider teaching scenes. The coming of education big data era brings not only opportunities for the development of EDM, but also challenges in technology and management. In the future, it will be the research trend of EDM in the era of big data of education in the future to integrate the relevant knowledge of education field and the experience of teachers’ teaching process into the established model. We expect it to be more mature and prosperous.

References 1. Xu, P., Wang, Y., Liu, Y.: Analysis of learning change from the perspective of big data: interpretation and enlightenment of the report “promoting teaching and learning through education data mining and learning analysis” in the United States. J. Distance Educ. 11 (2013) 2. Sun, Q., Lu, C.: Research on the application of big data in colleges and universities. China Educ. Netw. 63–65 (2014) 3. Chen, D., Zhan, Y., Yang, B.: Application analysis of deep learning technology in education big data mining. Audio Vis. Educ. Res. 40(02), 70–78 (2019) 4. Bienkowski, M., Feng, M., Means, B.: Enhancing teaching and learning through educational data mining and learning analytics: an issue brief. Technical report. Office of Educational Technology, U.S. Department of Education, Washington, pp. 1–57 (2012) 5. Baker, R.S.J.D., Yacef, K.: The state of educational data mining in 2009: a review and future visions. J. Educ. Data Min. 1(1), 3–17(2009) 6. Merceron, A., Yacef, K.: Educational data mining: a case study. In: Conference on Artificial Intelligence in Education: Supporting Learning Through Intelligent & Socially Informed Technology (2005) 7. Romero, C., Ventura, S.: Educational data mining: a survey from 1995 to 2005. Expert Syst. Appl. 33(1), 135–146 (2007) 8. Liu, F.: A review of the application of big data in education. Mod. Educ. Technol. 24(8), 13–19 (2014) 9. Chen, C., Wang, Y., Li, C.: Research and application of big data for online education. Comput. Res. Dev. 67–74 (2014) 10. Zhou, Q., Mou, C., Yang, D.: Summary of research progress in education data mining. J. Softw. 26(11), 3026–3042 (2015) 11. Yang, X., Wang, D.D.: Application mode and policy suggestions of education big data. Res. Audio Vis. Educ. 54–61 (2015)

278

X. He et al.

12. Chai, Y., Lei, C.: Overview of online learning behavior research based on data mining technology. Comput. Appl. Res. 1287–1293 (2018) 13. Yu, F., Liu, Y.: “User centered” education data mining application research. Audio Vis. Educ. Res. 39(11), 69–77 (2018) 14. Sun, X., Wu, N., Zhang, L.: Prediction method of MOOCS dropout rate based on deep learning. Comput. Eng. Sci. 893–899 (2019) 15. Chaplot, D., Rhim, E., Kim, J.: Predicting student attrition in MOOCs using sentiment analysis and neural networks 16. Wang, X., Zou, G., Li, X.: Prediction of learners’ dropping out of class based on MOOC data. Mod. Educ. Technol. 95–101 (2017) 17. Xing, W., Du, D.: Dropout prediction in MOOCs: using deep learning for personalized intervention. J. Educ. Comput. Res. (2018) 18. Márquez-Vera, C., Cano, A., Romero, C.: Early dropout prediction using data mining: a case study with high school students. Expert Syst. 33(1), 107–124 (2016) 19. Lakkaraju, H., Aguiar, E., Shan, C.: A machine learning framework to identify students at risk of adverse academic outcomes. In: ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 1909–1918. ACM (2015) 20. Wu, J.: Literature review on collaborative filtering recommendation algorithm. Shang, p. 224 (2016) 21. Zhang, Y.: Research and design of personalized learning recommendation system based on LSTM (2018) 22. Yang, H.: Research on personalized recommendation method based on deep belief network in MOOC environment. Central China Normal University (2018) 23. Lu, X., Wang, X.: Mining association rules of academic data based on domain association redundancy. Comput. Sci. 427–430 (2019) 24. Luna, J.M., Romero, C., Romero, J.R., Ventura, S.: An evolutionary algorithm for the discovery of rare class association rules in learning management systems. Appl. Intell. 42(3), 501–513 (2014) 25. Geigle, C., Zhai, C.X.: Modeling MOOC student behavior with two-layer hidden Markov models. In: Fourth ACM Conference on Learning (2017) 26. Rabbany, R., Elatia, S., Takaffoli, M., Zaïane, O.R.: Collaborative learning of students in online discussion forums: a social network analysis perspective. In: Peña-Ayala, A. (ed.) Educational Data Mining. SCI, vol. 524, pp. 441–466. Springer, Cham (2014). https://doi. org/10.1007/978-3-319-02738-8_16 27. Saqr, M., Fors, U., Tedre, M., Nouri, J.: How Social Network Analysis Can Be Used to Monitor Online Collaborative Learning and Guide an Informed Intervention (2018)

Research on DINA Model in Online Education Jia Liu, Wensheng Tang(B) , Xuping He, Bo Yang, and Shengchun Wang College of Information Science and Engineering, Hunan Normal University, Changsha 410081, Hunan, People’s Republic of China [email protected], {tangws,yangbo}@hunnu.edu.cn, [email protected], [email protected]

Abstract. Learning tests play an important role in both traditional learning and online learning. The traditional education test can only report students’ scores or abilities, but not their knowledge level, which is no longer satisfied with people’s requirements. In recent years, DINA model of cognitive diagnosis model has been widely used to diagnose students’ knowledge mastery. DINA model can dig out the students’ knowledge points and give feedback to the teachers, so that the teachers can make remedial plans for the students’ deficiencies in time. This paper first introduces the basic principle of DINA model and the improvement of DINA model in the field of education in recent years. Secondly, we introduce the development of Dina model under the trend of online education, and prove the availability of DINA model in online platform with experimental data. Finally, we predict and analyze the research direction of DINA algorithm in the future. Keywords: Cognitive diagnosis · DINA model · Online education

1 Introduction As an important part of students’ learning, test plays an important role in traditional education and online learning. With the maturity of testing theory and its wide application, people have a higher pursuit of testing. In traditional education tests, only the test scores or ability values of the testes can be reported and taken as the evaluation index. In fact, the subjects with the same score have different cognitive structures and different cognitive processes [1]. In 2001, the United States passed the bill “No Child Left Behind Act of 2001”. The act requires that each student’s diagnostic information be added to the educational test report. Research shows that education with only one test result but This work was supported in part by the Hunan Province’s Strategic and Emerging Industrial Projects under Grant 2018GK4035, in part by the Hunan Province’s Changsha Zhuzhou Xiangtan National Independent Innovation Demonstration Zone projects under Grant 2017XK2058, in part by the National Natural Science Foundation of China under Grant 61602171, in part by the Scientific Research Fund of Hunan Provincial Education Department under Grant 17C0960 and 18B037, and in part by the Key Research and Development Program of Hunan Province under Grant 2019SK2161. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 279–291, 2020. https://doi.org/10.1007/978-3-030-63955-6_24

280

J. Liu et al.

no diagnosis and remedy is meaningless. Now people need not only the superficial evaluation of ability, but also the internal cognition [2]. Therefore, modern psychological researchers combined cognitive psychology and psychometrics to develop a new psychometric model with cognitive diagnosis function, namely cognitive diagnosis model (CDM) [3]. In cognitive psychology, cognitive diagnosis model can be used to model students’ cognitive state from the knowledge point level. After the theory came into being, it attracted wide attention. After the model of cognitive diagnosis came into being, it has attracted a lot of attention. Common cognitive diagnostic models include: Linear logistic trait model proposed by Fisher [4]; Rule space model proposed by Tatsuoka et al. [5]; Hartz’s fusion model [6]; attribute hierarchy mode by Leighton [7]; the deterministic input noise and gate model (DINA) proposed by Torre et al. [8]; a neural network-based PSP method proposed by Qian et al. [9] and so on. According to its theoretical basis, cognitive diagnosis models are divided into four categories in literature [1]: (1) potential trait models; (2) methods based on non-parametric artificial intelligence; (3) potential classification models; (4) evidence center design. Among the cognitive diagnostic models, Dina model is one of the most popular. The model has simple parameters and wide application range. It can dig out the students’ knowledge points and give feedback to the teachers, so that the teachers can make remedial plans for the students’ deficiencies in time. This article studies the DINA model, focusing on the application of the DINA model in the context of new information technology. The logic structure of this paper is shown in Fig. 1, and the basic structure is as follows: The second section introduces the principle of DINA model; the third section introduces the application of DINA model under online education, including fuzzy diagnosis, common diagnosis and large sample diagnosis; the fourth section makes experimental analysis on the availability of DINA model under online education, and compares it with the traditional algorithm; the fifth section puts forward the prospect.

DINA Model Basics

Usability Analysis of DINA Model

Application of DINA Model in Online Education Large Sample

Fuzzy Diagnosis

Diagnosis

General Character Diagnosis

Fig. 1. Basic framework of this study

Research on DINA Model in Online Education

281

2 Model Specification The DINA model is a typical discrete cognitive diagnostic model. This model describes the students as a multi-dimensional knowledge point grasping vector, and diagnoses from the actual answer results of the students. It can determine whether the test individual has mastered each skill required to correctly answer the test questions, help students understand their deficiencies. And let teachers teach purposefully [10]. In the DINA model, a student’s knowledge vector and the Q-matrix constitute a latent response variable for student m to question v: ηmv =

K k=1

q

αmkvk .

(1)

Here, ηmv is a potential response indicator, ηmv = 1 means the student Im mastered the test question Jv . If not, ηmv = 0. K is the number of attributes. Let αmk be the binary variable for student i on knowledge k(k = 1, . . . , K), where a value of 1 indicates that student m shows mastery of knowledge k and a value of 0 indicates non-mastery, and let αmk be the vector of knowledge of student m. Let Q be a V by K matrix, with element qvk indicating whether knowledge k is required to answer question j correctly. The probability that student m correctly answers question v(Ymv = 1) is expressed in the DINA model as follows: P(Ymv |αm ) = gv1−ηmv (1 − sv )ηmv ,

(2)

Where the slipping parameter sj is the probability of an incorrect response to question m if all of the required knowledge of question j have been mastered, that is, sj = P(Ymv = 0|ηmv = 1), and the guessing parameter gj is the probability of a correct response to question m if a student lacks at least one of the required knowledge for question m, that is, gj = P(Ymv = 1|ηmv = 0). The parameter estimation methods commonly used in cognitive diagnostic models are the maximum expectation algorithm (Expectation Maximization Algorithm, EM) [8] and Markov Chain Monte Carlo Algorithm (MCMC) [11]. MCMC algorithm is more suitable for complex multi-parameter estimation. In the DINA basic model, only the sj and gj needs to be calculated by parameter estimation. We use the EM algorithm for parameter estimation. EM algorithm is an iterative algorithm for finding maximum likelihood estimates of parameters in a probability model. It has two steps: the first step is Expectation Step, which is called the E step, and the second step is the maximization step, which is called the M step. Each iteration can ensure that the value of the likelihood function increases and converges to a local maximum. For detailed steps, please refer to the literature [8]. The DINA model uses the EM algorithm to maximize the edge likelihood of (2) to obtain parameter estimates. The likelihood function is expressed as: L(s, g|α) =

I m=1

J v=1

mv

PV (αm )Ymv [1 − Pv (αm )]1−Y .

(3)

The diagnostic process of the DINA model is shown in Fig. 2. As can be seen from Fig. 2, the diagnostic process mainly includes the input of candidates’ answers, the diagnosis of mastery, and the output of score prediction. Before diagnosis, the Q matrix is defined

282

J. Liu et al.

by an expert and is used as a known condition input with the student’s answer matrix Y to diagnose and obtain the student’s potential master set S. Then, the master can be used to diagnose and predict the student’s response to the unanswered questions. Feedback

Prediction score matrix y

test

prediction

Question score matrix y

Cognitive

Knowledge

Exam Questions Knowledge Points

Fig. 2. Cognitive diagnosis process

In the education field, improvements have been made by many scholars to the DINA model, mainly divided into the following three categories: (1) Improvement of Q matrix. The Q matrix in the traditional DINA model is a subjective process and model parameter estimates may be inaccurate in cases where researchers employ an unspecified Q. Some studies assume part of the knowledge in the Q matrix, and have developed refining programs and algorithms. For example, de la Torre and Chiu provide a validation process that first estimates the project parameters in a temporary Q matrix, and then updates the elements of Q with an algorithm [12]. A fundamentally different approach to estimation of Q is introduced in Chen et al. [13]. This technique provides a prior specification of Q, which can be extracted from the posterior distribution by MCMC algorithm. (2) Multi-strategy improvements. When examinees can apply multiple strategies to solve problems, single strategy model (such as SS-DINA model) cannot fully capture the complexity of multiple strategy features [14]. For test tasks with multiple strategies, different strategies may be applied by different students [15]. Multi Strategies diagnosis model is an inevitable choice. For example, De la Torre and Dougllas proposed MS-DINA model to provide multiple solutions to problems [15]. The model establishes multiple q-matrices, corresponding to different strategies, and students only need to use one of the strategies to solve the questions, which is more suitable for the actual situation of students. (3) Multi-level scoring improvements. Almost all cognitive diagnostic models can only adapt to binary data, which limits the application and development of cognitive diagnostics and cannot meet the needs of practical work [3]. Tu, Cai, et al. combined the idea of cumulative category response function in the hierarchical response model proposed by Samejima, and extended the dichotomy DINA model to a multi-model

Research on DINA Model in Online Education

283

called P-DINA model [3]. In addition, Torre relaxed the constraints and proposed a generalized Dina model [16]. With the development of online education, cognitive diagnosis is no longer confined to traditional classrooms, and researchers have begun to develop online platforms.

3 Development of the DINA Model in Online Education With the continuous development of information science, online education platforms such as MOOC and Netease cloud classroom are springing up rapidly. As an important supplement and extension of the traditional education mode, online education can provide a novel teaching method. However, at present, existing algorithms for analyzing students’ learning situation are basically based on line test questions, which are not based on students’ knowledge points and are prone to produce large deviation. In traditional learning, Torre puts forward a DINA model, which can get students’ knowledge points by question answering. If DINA model is applied to online education platforms to diagnose students’ learning situation, it will be helpful for online education. Therefore, scholars have developed a variety of application methods of DINA model on online education platforms. These methods have great practical value in using DINA model from different entry points. 3.1 Fuzzy Diagnosis The diagnosis of students’ learning situation in online learning includes two parts: objective problem diagnosis and subjective problem diagnosis. Although the traditional DINA model for the diagnosis of objective questions (only two kinds of answers to the wrong questions) has been gradually matured, but it is not very good to diagnose the subjective questions. In order to apply DINA model to online education platforms, it is first necessary to break the limitation that DINA model can only diagnose objective questions, so that it can diagnose subjective questions. After mastering the learning situation of students, online education platform can carry out personalized teaching for students. Most people’s understanding of knowledge is based on vague concepts. In the data source, it is very difficult to accurately find the relationship between different value points or value ranges of each attribute, and people are more concerned about the high level of abstraction [17]. In 1965, Zadeh established a method, fuzzy theory, to describe fuzzy phenomena on the basis of mathematics [18]. At present, fuzzy set theory has been successfully applied in many fields such as pattern recognition, intelligent control, machine learning, and artificial intelligence [19]. Like the traditional set operation, in the fuzzy theory, fuzzy sets also have fuzzy intersection and fuzzy union calculation - fuzzy Union takes the larger one, and fuzzy intersection takes the smaller one. The formula is: A ∩ B(x) = min(A(x), B(x)).

(4)

A ∪ B(x) = max(A(x), B(x)).

(5)

284

J. Liu et al.

Wu et al. propose a fuzzy cognitive diagnosis framework (FuzzyCDF) for examinees’ cognitive modelling with both objective and subjective problems [20]. In this model, the probability expressions of students answering subjective and subjective questions are:       (6) P Xij = 1|ηij , sj , gj = 1 − sj ηij + gj 1 − ηij .       P Xij = 1|ηij , sj , gj = N (Xij | 1 − sj ηij + gj 1 − ηij , σ 2 ).

(7)

Here, (6) is the diagnosis of objective problems. There are only two answers for students, either right or wrong, so Xij takes the value 0 or 1. In (7), students have many possibilities to answer subjective questions. So Xij is transformed into a continuous value of [0, 1]. σ is the standardized variance of the subjective question score. The degree αik of the knowledge point in the model, that is, the degree of membership i(k) between the student i and the knowledge point k, is determined by the degree of membership of the fuzzy set related to the knowledge point. Formally, αik is defined as: αik = i(k) =

1 , 1 + exp[D ∗ αik (θi − bik )]

(8)

the meaning of this definition is that the candidate’s proficiency in a particular skill(αik ) depends on the difference between the candidate’s advanced trait (θi ) and the attributes of the skill: difficulty (bik ) and discrimination (αik ) of skill k for student j. Here, the coefficient D is set to 1.7, which is an empirical scale constant in a log-cognitive model. ηij is student i‘s mastery of test j. In objective questions, it is defined as the fuzzy intersection of student i‘s grasp of all the knowledge points examined in test question j, expressed as ηij = i ∩i≤k≤K,qk=1 (k) ; in subjective questions, it is the fuzzy union of student i’s grasp of all the knowledge points examined in test j, which is expressed as ηij = i ∪i≤k≤K,qk=1 (k) . Fuzzy-CDF can extract information from objective and subjective issues, obtain more accurate and interpretable cognitive analysis, and analyze the characteristics of each student, which has great practical effects. In 2016, Liu et al. proposed the SDINA model [21] (soft deterministic inputs, noisy “and” gate model Model), which considers the posterior probability of each knowledge point, improves the parameter estimation process of Dina model, and changes the knowledge point mastery of students into a continuous value between 0 and 1, making the diagnosis result more accurate. Later, some researchers proposed R-FuzzyCDF model based on FuzzyCDF [22]. This model redefines the mastery degree of knowledge points in the FuzzyCDF model and the mastery degree of subjective test questions by students. R-FuzzyCDF model relates the importance of knowledge points with the number of subsequent knowledge points and the number of related questions, that is, the more a knowledge point is used by other knowledge points, the more important the knowledge point is. In addition, with the increase of the number of knowledge points mastered by students, the probability of a correct answer will increase. To some extent, R-FuzzyCDF model makes up for the deficiency of FuzzyCDF, which further improves the accuracy of diagnosis model. But the parameters to be estimated also become more, which to some extent destroys the feature of simple parameters of DINA model.

Research on DINA Model in Online Education

285

3.2 General Character Diagnosis The DINA model can diagnose students’ knowledge mastery according to the students’ questions, and then contact the students ‘knowledge mastery to evaluate the students. In this way, students’ weak learning parts can be found and consolidated. However, the DINA model can only obtain diagnostic results based on its own conditions, without considering the commonality between students. It is too one-sided, and the results may have large errors, so it has not been promoted and has not played its true role. In the development of e-commerce, how to recommend favorite products to users of online malls is a hot topic. To accurately recommend products to users, it is necessary to first analyze and diagnose the users, extract the characteristics of the users, and then judge the users based on the characteristics. Collaborative filtering algorithms are the most popular type of recommendation system in applied research. It takes into account the performance of similar users to model the target user. When the collaborative filtering algorithm is applied to student diagnosis, it takes into account the commonalities among the students, instead of unilaterally considering the target students. However, because collaborative filtering algorithms cannot analyze the students themselves, they are not interpretable. Researchers combine the DINA model with collaborative filtering algorithms and use their respective strengths to make up for each other’s shortcomings to achieve the effect of common diagnosis. Zhu [23] and others combined the DINA model and the probability matrix decomposition method in collaborative filtering to propose a PMF-CD model. In the calculation process, the knowledge level of the students given by the previous DINA model was only 1 and 0. The PMF-CD model considers all posterior probabilities of the degree of mastery of the knowledge points, and converts the degree of mastery of the knowledge points into continuous values between 0 and 1. Then, they modeled how well they mastered the test questions based on their level of knowledge. Finally, the student’s true answer level is used as the prior information of the probability matrix decomposition algorithm, and the student’s score matrix is decomposed into the student’s feature matrix and the test question’s feature matrix to calculate the student’s potential correct answer probability. Compared with a single DINA model, the PMD-CD model considers the connection between students and is more interpretable than the DINA model. It takes into account the learning personality between students and the commonality of student learning, and can more effectively analyze the learning of students. Based on the probability matrix decomposition method, researchers have successively combined other algorithms of collaborative filtering with cognitive diagnosis to obtain the best diagnostic results. Qi B proposed a test recommendation method based on collaborative filtering and cognitive diagnosis in an adaptive testing environment, extended the application rules of the DINA model on multi-level attribute scoring, and expanded the application scenarios of test recommendation [24]. Reference [25] takes into consideration the knowledge of students ‘knowledge points and recommends secondary collaborative filtering test questions based on knowledge points, and then uses item response theory and deep self-encoder to predict students’ fullness of recommended knowledge points on recommended test questions and Comprehensive and wonderful, and finally generate a list of recommended test questions based on the prediction results.

286

J. Liu et al.

Shan R T combines the SVD algorithm with the DINA model [26]. When analyzing target students, the algorithm adds similar user performance to the target student diagnosis. Regarding the wrong question raised by the target user, it is believed that the target user has not grasped the relevant knowledge points of the problem. For the question raised by the target user, if similar users made mistakes, it will be also thought that the target user may not be able to grasp the relevant knowledge points of the problem. Diagnostic results can further reduce calculation errors. Li modeled the students through the cognitive diagnosis model [27], and then integrated the knowledge points of the test questions and the students’ knowledge points into the matrix according to the joint probability matrix decomposition method. This method is a good description of the performance of students and test questions, ensuring the interpretability and rationality of scores. Moreover, through time complexity analysis, the method can be extended to large-scale data sets and applied to online education platforms for big data. 3.3 Large Sample Diagnosis With the development of Internet education, educational methods have also undergone major changes with changes in computer technology and people’s teaching concepts [28]. A large number of students have begun to use online education platforms for learning. Faced with such huge and complicated data on online education platforms, the research and application of big data in the field of online education are more important [29]. The DINA model is a good diagnostic model. In recent years, researchers have applied it to online education. However, traditional DINA models are diagnosed by students with less sample data. When faced with an online education platform for large-scale student groups, the diagnostic efficiency of the DINA model will be greatly reduced [30], which makes the application of the DINA model very limited. Therefore, how to optimize the diagnostic efficiency of the DINA model has always been the focus of scholars’ research. In 2004, in order to reduce the time complexity of the DINA model and improve its efficiency, Torre proposed the HO-DINA model [11]. This model assumes that the control of knowledge points is controlled by a hyperparameter to reduce the dimensions of knowledge points. Although this method can improve the calculation efficiency to a certain extent, it destroys the good interpretability of the original DINA model. In the DINA model, there are hidden variables that cannot be directly observed, that is, the degree of mastery of the students’ knowledge points, so the EM algorithm (see the introduction in Sect. 2) is needed for parameter estimation. However, when the amount of missing information is large, the calculation speed of the EM algorithm becomes slower [31], which will reduce the diagnostic efficiency of the DINA model. In order to improve efficiency while maintaining the characteristics of the DINA model, Wang C et al. started from the EM algorithm used for parameter estimation in the DINA model and indirectly achieved the effect of diagnostic acceleration by improving the calculation efficiency of the EM algorithm [32]. The acceleration of the DINA model is achieved by increasing the parameter estimation speed through the following three different data set partitioning methods: 1) Incremental DINA model (I-DINA). The model mainly divides the student data, and accesses only one student block during each iteration to update the likelihood function. Other students who have not yet visited retain the results of the last iteration. 2) Maximum Entropy DINA model

Research on DINA Model in Online Education

287

(ME-DINA). The ME-DINA model filters out data sets that have little effect on each iteration to form a lazy set. The changed data set is called a change set. During the next iteration, the change set will be iterated and the lazy set will not be iterated. The final result will be retained. 3) A hybrid model based on the former two—Incremental Maximum Entropy DINA (IME-DINA). In the change set of the ME-DINA model, the division of student blocks was added. Experiments show that all three methods can achieve acceleration effects of several to several tens of times while ensuring the validity of the DINA model, which effectively improves the calculation efficiency of the DINA model. I-DINA calculation is fast, and ME-DINA calculation is stable while IME-DINA model combines the advantages of both and can achieve stable and fast results. Moreover, the above three methods are not limited to the DINA model and can also be continuously applied to other cognitive diagnostic models based on the EM algorithm, which have good generality and promotion value.

4 Usability Analysis of DINA Model In online education, it is important to predict the next answer of a student based on their learning and recommend personalized test questions for the student. This article takes the test question recommendation algorithm as an example. The traditional test recommendation algorithm is compared with the recommendation algorithm added to the DINA model, and the availability of DINA is proved by experimental results. 4.1 Dataset Introduction The data set comes from reference [20], which are FrcSub data set, Math1 data set and Math2 data set. Table 1 shows the statistics of the three data sets. The three data sets contain two data, matrix Q and scoring matrix X. Each column of matrix Q represents a knowledge point, and each row represents a test problem. 0 means the test question does not contain the knowledge point, and 1 means it does. Each column of score matrix X represents the score of a test question, each row represents a test question, 0 indicates that the student’s score on the question is 0, and 1 indicates that the student’s score is 1. Table 1. Relevant statistics of data set Student Questions Knowledge FrcSub

536

20

8

Math1

4209

20

11

Math12 3911

20

16

288

J. Liu et al.

4.2 Experimental Evaluation Indicators This article uses the precision, recall and F1 indicators commonly used in recommendation fields [24]. The specific definitions are as follows: precision = recall = F1 =

TP . TP + FP

TP . TP + FN

2 ∗ precision ∗ recall . precision + recall

(9) (10) (11)

Here, TP indicates the number of students who really answered correctly in the recommended test questions, FP indicates the number of students who incorrectly answered the recommended test questions, and FN indicates the number of students who really answered the recommended test questions. 4.3 Comparative Test Methods (1) DINA model [8]. Calculate students ‘knowledge of mastery according to the DINA model, and select students’ weak knowledge points from test questions for recommendation. (2) User-based collaborative filtering [33]. This method selects similar students to do wrong test questions to recommend to users. (3) Collaborative filtering algorithm based on the DINA model. Comprehensive consideration of user-based collaborative filtering algorithm and DINA model. Choosing test questions with weak students ‘knowledge points and similar students’ mistakes and recommend them. 4.4 Analysis of Experimental Results Select 30%, 40% and 50% of the test questions as the test set, recommend three test questions to the students of the test set, and calculate the accuracy, recall, and F1 values of the three methods, respectively. The experimental results are shown in Table 2. From this figure, we have the following conclusions: (1) Observed from the data, the collaborative filtering algorithm that takes into account the degree of knowledge of the students’ knowledge is better than the traditional cognitive diagnostic model, and recommends to the students based on the knowledge of the similar users of the target user Title, the algorithm is more reliable and persuasive. (2) When the data is larger, the performance of the improved algorithm based on cognitive diagnosis collaborative filtering algorithm is superior, which shows that the algorithm is more suitable after adding DINA model, and can be applied to real life.

Research on DINA Model in Online Education

289

Table 2. Accuracy, recall and F1 of different recommendation methods in the same dataset Method

FrcSub 30%

Math1 40%

50%

30%

Math2 40%

50%

30%

40%

50%

Accuracy (1)

0.4264 0.4336 0.4421 0.3535 0.3323 0.3110 0.4208 0.3833 0.3693

(2)

0.8723 0.8729 0.8535 0.6735 0.6479 0.6470 0.6017 0.6183 0.6052

(3)

0.9067 0.8852 0.8744 0.7132 0.7200 0.7151 0.8042 0.7698 0.7348

(1)

0.4617 0.3365 0.2799 0.3679 0.2592 0.1920 0.4117 0.2816 0.2161

(2)

0.3365 0.3157 0.2937 0.1657 0.1674 0.1084 0.2777 0.2786 0.2692

(3)

0.3659 0.2611 0.2165 0.2766 0.3097 0.2848 0.2469 0.2048 0.1828

(1)

0.4433 0.3789 0.3428 0.3606 0.2913 0.2381 0.4162 0.3246 0.2727

(2)

0.4546 0.4362 0.4029 0.2660 0.1843 0.1392 0.3021 0.3147 0.2882

(3)

0.5214 0.4032 0.3471 0.3773 0.3834 0.3549 0.3778 0.3359 0.2927

Recall

F1

5 Conclusion Cognitive diagnosis models are bridges between internal attribute and external response, which plays an important role in cognitive diagnosis and evaluation [34]. It can accurately measure and diagnose the cognitive attribute structure of students, provide exact basis for improving or remedying teaching, and point out the direction for improving teaching quality [35]. Compared with other models, DINA model is an excellent model with simple parameters and easy to understand [34], which has been widely used in the field of education. Researchers have also constantly improved the model. In recent years, with the development of social information, education has been transforming towards educational information. The rapid rise of various online education platforms will be a major development trend of education in the future. In this development trend, the application of DINA model to online education platforms can help to analyze the status of students, thus contributing to the development of online education. Therefore, researchers continue to integrate DINA model and information technology, so that DINA model can get rid of the limitations of traditional education. Although DINA model has developed for many years, it is still in its infancy in information education. In the future, our research on DINA model can be carried out from the following aspects: (1) How to show the diagnosis process of students? Hidden attributes of students are difficult to measure in DINA model. If the diagnosis process is shown in the diagnosis, the reliability of the diagnosis will be greatly improved. (2) How to explore other learning factors of students? In the DINA model, only two factors of students’ guesses and errors are considered. In fact, students are influenced by many factors when they complete a set of test questions. Fully mining the potential factors of students can help teachers better understand students’ learning state and help improve students’ academic level.

290

J. Liu et al.

(3) How to make cognitive diagnosis of multi-disciplinary integration? Current cognitive diagnosis work is all around the test of a single subject for students. However, students learn multiple subjects at the same time, and the subjects affect each other, actually. Through the combination of data mining technology and cognitive diagnosis, the learning situation of students can be directly understood through students’ multi-disciplinary integration for diagnosis. (4) How to overcome the data redundancy brought by big data in the diagnosis process? With the development of the era of big data, how to overcome the data redundancy and recommend the students’ performance prediction quickly and stably in the massive data is the next issue to be considered.

References 1. Chen, Q.M., Zhang, M.Q.: Development of cognitive diagnosis models and their application methods. Adv. Psychol. Sci. 18(03), 522–529 (2010) 2. Dongbo, T., Cai, Y., Qi, S., Dai, H.: A review on cognitive diagnostic models under modern test theory. Psychol. Explor. 28(2), 64–68 (2008) 3. Tu, D.B., Cai, Y., Dai, H.-Q.D.: A polytomous cognitive diagnosis model: P-DINA model. Acta Psychologica Sinica 42(10), 1011–1020 (2010) 4. Fischer, G.H.: The linear logistic test model as an instrument in educational research. Acta Physiol. (Oxf) 7(6), 359–374 (1973) 5. Tatsuoka, K.K.: Rule space: an approach for dealing with misconceptions based on item response theory. J. Educ. Meas. 20(4), 345–354 (1983) 6. Hartz, M.C.: A Bayesian framework for the unified model for assessing cognitive abilities: blending theory with practicality. Am. J. Gastroenterol. 95(4), 906–909 (2002) 7. Leighton, J.P., Gierl, M.J., Hunka, S.M.: The attribute hierarchy method for cognitive assessment: a variation on Tatsuoka’s rule-space approach. J. Educ. Meas. 41(3), 205–237 (2004) 8. De La Torre, J.: DINA model and parameter estimation: a didactic. J. Educ. Behav. Stat. 34(1), 115–130 (2009) 9. Qian, J.X., Yu, J.Y.: The neural network-based PSP method in cognitive diagnosis. Psychol. Sci. 33(4), 915–917 (2010) 10. Ozaki, K.: DINA models for multiple-choice items with few parameters: considering incorrect answers. Appl. Psychol. Meas. 39(6), 431–447 (2015) 11. De La Torre, J., Douglas, J.A.: Higher-order latent trait models for cognitive diagnosis. Psychometrika 69(3), 333–353 (2004) 12. de la Torre, J., Chiu, C.Y.: A general method of empirical Q-matrix validation. Psychometrika 81(2), 253–273 (2016) 13. Chen, Y., Culpepper, S.A., Chen, Y., Douglas, J.: Bayesian estimation of the DINA Q matrix. Psychometrika 83(1), 89–108 (2018) 14. Liu, T.H., Zhao, Y., Dai, H.: Mix-DINA model based on mixture distribution for multiple strategy cognitive diagnosis. Psychol. Explor. 36(5), 464–471 (2016) 15. Huo, Y., de la Torre, J.: Estimating a cognitive diagnostic model for multiple strategies via the EM algorithm. Appl. Psychol. Meas. 38(6), 464–485 (2014) 16. De La Torre, J.: The generalized DINA model framework. Psychometrika 76(2), 179–199 (2011) 17. Xiong, X.H., Yao, J.C.: Research and application of data mining based on fuzzy sets. Comput. Eng. Appl. 01, 203–205 (2002)

Research on DINA Model in Online Education

291

18. Zadeh, L.A.: Fuzzy sets. Inf. Control 8(3), 338–353 (1965) 19. Li, G., Yang, B., Liu, Y., et al.: Survey of data mining based on fuzzy set theory. Comput. Eng. Des. 32(12), 4064–4067 (2011) 20. Wu, R., Liu, Q., Liu, Y., et al.: Cognitive modelling for predicting examinee performance. In: International Conference on Artificial Intelligence, pp. 1017–1024. AAAI Press (2015) 21. Liu, Y., et al.: Collaborative learning team formation: a cognitive modeling perspective. In: Navathe, S.B., Wu, W., Shekhar, S., Du, X., Wang, X.S., Xiong, H. (eds.) DASFAA 2016. LNCS, vol. 9643, pp. 383–400. Springer, Cham (2016). https://doi.org/10.1007/978-3-31932049-6_24 22. Li, Y.X., Wen, Y.M., Yi, X.H., et al.: Revised model of fuzzy cognitive diagnosis framework. J. Data Acquisition Process. 05, 110–121 (2017) 23. Tianyu, Z., Zhenya, H., Enhong, C., et al.: Personalized test item recommendation method based on cognitive diagnosis. Chin. J. Comput. 01, 178–193 (2017) 24. Qi, B., Zou, H.X., Wang, Y., et al.: Question recommendation based on collaborative filtering and cognitive diagnosis. Comput. Sci. 235–240 (2019) 25. Xiong, H.J., Song, Y.F., Zhang, P., et al.: Personalized question recommendation based on autoencoder and two-step collaborative filtering. Comput. Sci. 172–177 (2019) 26. Shan, R.T., Luo, Y.C., Sun, Y.: Collaborative filtering algorithm based on cognitive diagnosis. Comput. Syst. Appl. 27(3), 136–142 (2018) 27. Li, Q., Liu, X.H., Xu, X.H., et al.: Personalized test question recommendation method based on unified probalilistic matrix factorization. J. Comput. Appl. 38(3), 639–643 (2018) 28. Aparicio, M., Bacao, F.: E-learning concept trends. In: Proceedings of the 2013 International Conference on Information Systems and Design of Communication, pp. 81–86. ACM, New York (2013) 29. Chen, C., Wang, Y.P., Li, C., et al.: The research and application of big data in the field of online education. J. Comput. Res. Dev. S1, 67–74 (2014) 30. Torre, J., Minchen, N.: Cognitively diagnostic assessments and the cognitive diagnosis model framework. Educ. Psychol. 20(2), 89–97 (2014) 31. Dempster, A.P., Laird, N.M., Rubin, D.B.: Maximum likelihood from incomplete data via the EM algorithm. J. Roy. Stat. Soc.: Ser. B (Methodol.) 39(1), 1–22 (1977) 32. Wang, C., Liu, Q., Chen, E.H., et al.: The rapid calculation method of DINA model for large scale cognitive diagnosis. Acta Electronica Sinica 46(5), 1047–1055 (2018) 33. Goldberg, D., Nichols, D., Oki, B.M., et al.: Using collaborative filtering to weave an information tapestry. Commun. ACM 35(12), 61–70 (1992) 34. Xiao, Z., Ruxue, S.: Research advance in DINA model of cognitive diagnosis. China Examinations 1, 32–37 (2013) 35. Zhang, M.Q., Fan, X.Z., Guo, K.Y., et al.: The evolution and development of education quality assessment under the concept of big data. Mod. Educ. J. 3, 2–6 (2016)

Design of Data-Driven Visualization Teaching System for Preschool Basketball Courses Y.-f. Qin1 , L.-z. Mo2(B) , and Can Chen3 1 Department Sports and Education, Guang Xi College of Physical Education,

Nanning 530012, China 2 Nanning University, Nanning 530200, China

[email protected] 3 Nursing Home of the Committee of the Guangxi Zhuang Autonomous

Region of the Communist Party, Nanning 530022, China

Abstract. Some studies have shown that basketball can promote the development of young children’s nervous system, exercise their physical coordination ability, allow young children to have a team awareness, increase young children’s social time, and establish social relationships. However, there is a problem of high system resource consumption level in the traditional visualized teaching system for infant basketball courses. Therefore, a data-driven visualized teaching system design for infant basketball courses is proposed. The hardware part adopts the intelligent display terminal DMT80480507OSW to realize human-computer interaction, which allows children to increase learning interest through human-computer interaction. The software part uses data compression algorithms to compress basketball course resource data. After testing, the feasibility of the design system was proved to provide richer teaching resources for children’s basketball courses. Keywords: Data driven · Infant basketball · Visualization · Teaching system

1 Introduction With the continuous development and progress of the country and society, the value and significance of preschool physical education have been increasingly affirmed and valued. Researchers have gradually realized that by paying attention to whether physical education in educational institutions provides young children with opportunities for comprehensive physical and mental development, including opportunities for physical and psychological development [1]. As the first stage of the education system, kindergartens have assumed more responsibilities. It is here that children’s lack of activities caused by the restrictions of civilization is first compensated; it is also here that children get a living space that meets their needs [2]. Sports education has been considered as a field of experience established from a professional perspective, and it is formally represented as “sports lessons” that are regularly conducted in kindergartens.

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 292–302, 2020. https://doi.org/10.1007/978-3-030-63955-6_25

Design of Data-Driven Visualization Teaching System

293

Ball games have a broad mass base in China, especially with the continuous improvement of people’s living standards, the development needs of national sports, and the frequent appearance of Chinese ball players in the news media in high-level competitions in the world, such as Yao Ming in basketball, Li Na and Zheng Jie in tennis, Zhang Jike and Zhang Yining in table tennis, etc. To make ball games a popular sport among educational institutions, families, communities, especially young children [3]. Studies have found that basketball has a good comprehensive effect on fitness and intelligence for young children. First, when children are exposed to basketball, they observe the direction, height, speed, distance, and size of the ball to promote the development of their perception of orientation, space, and time. Secondly, children’s brain, eyes, and limb movements coordinate and cooperate in the action of shooting, striking, and throwing. Judging the direction of the movement through the shouts of teachers and teammates can effectively promote the nervous system’s ability to control limb movements, as well as vision, hearing, The coordination and coordination of consciousness can improve the development of children’s motor skills [4]. Thirdly, the racquets and dribbling movements in basketball require the cooperation of left and right hands and feet. The coordinated movement of both sides of the limbs is conducive to improving the balanced development of left and right brain intelligence of children and laying a good foundation for the development of children’s learning and thinking skills [5] foundation. In addition, basketball requires teamwork, cooperation in relay, cooperation in offense and defense, and observance of the rules of the game, which is conducive to the benign shape of individual child socialization [6]. At the same time, in basketball, children can naturally recognize the rules of playing basketball, and use various thinking activities to win basketball games, which has promoted the intellectual development of children. In recent years, information technology has achieved remarkable results in all walks of life and has greatly promoted the progress and development of society. In the education industry, with the popularization of multimedia teaching and application updates, the advantages of information technology in teaching have become increasingly prominent, and it has continued to move beyond the pure application model to the independent research and development model, especially in higher education. The “informationdevelopment” combined education informationization development model is constantly emerging [7]. Practice has proved that the “Internet + education” education model is leading our country to the era of educational informatization 2.0, and using education to modernize education will be the only way to achieve high-quality and balanced development of education. Data-driven technology originates from the computer field, which is mainly guided by the data in the database during programming [8]. Starting from the existing problem facts, based on a large amount of intermediate data, combined with data processing and analysis methods, effective information is obtained from a large number of raw data to achieve multiple goals of data forecasting, data evaluation, data monitoring and diagnosis [9]. Data-driven technology can be used to make up for the shortcomings of traditional education methods. Designing a data-driven visualized teaching system for children’s basketball courses can also better solve the problem of excessive system resource consumption in the traditional system.

294

Y. Qin et al.

2 Overall Framework Design of Visualized Teaching System for Infant Basketball Courses The system architecture design uses the B/S architecture and uses ASP.Net technology to develop and design the infant basketball teaching system. Its characteristics are simple deployment and convenient maintenance. According to the project requirements and the mature design architecture on the current network, the entire system software architecture design uses the SSH framework [10]. The Web server is deployed using the IIS server, and the database is managed using the SQL database management system. The UI, business logic handler, and data access logic are all deployed by the Web server. The data access and control program uses.PHP technology to achieve connection and access to the SQL Server database. The user interface uses technologies such as HTML and JAVA to design and implement the UI interface. The UI interface interaction is implemented using relevant components of ASP.Net. The data structure, business thinking logic, and view presentation under this system are relatively separate and concrete, which is more conducive to system maintainability and scalability. The overall system architecture is shown in Fig. 1.

Client

Web server HTML ASP.NET

Data persistence layer

SQL Server

Javascript Javascript

Fig. 1. System architecture design

The whole system architecture combines the practical application mode, which is suitable for children’s learning characteristics. At the same time, it summarizes the valuable experience of B/S architecture time development, which has high reusability. As shown in the figure, according to J2EE development standard framework, the architecture of infant basketball teaching system includes three specific layers: presentation layer, business layer and data persistence layer. In practical application, combined with the automatic code generation tool, it supports the whole system architecture, and can also generate a large number of repetitive basic codes in a short time, which can make the practical development of early childhood teaching system focus on the development and design of business logic, effectively reduce the resource consumption of some links in the system development, and improve the quality and construction efficiency of the system Rise. The system architecture is described as follows: (1)Interface layer The interface mainly provides an interactive display interface for preschool basketball teaching systems and preschool teacher users. It is located at the client end of the system software and initially verifies the data on the display interface. In use, a request for a child basketball teaching service is sent from a Web browser, and then received by a

Design of Data-Driven Visualization Teaching System

295

Web server. The request information is encapsulated by the user, processed by the Web server, and then displayed in the user’s browser in a specified manner. The interface layer provides users with a unified data request entrance, which separates page display and data content. It is an international standard mechanism in modern information technology. It can form a consistent treatment of abnormal situations, is flexible, simple and effective, and facilitates function expansion and maintenance. (2)Business layer The business layer transforms the outgoing data structure of the interface layer into the data of the business layer and processes it. It is responsible for the logical thinking processing of various business work. After the system runs, it first accepts the user’s request, obtains the user’s basketball teaching data from the request data, calculates and processes it according to the specific rules, interacts through the way of components and services, and finally uploads the processing results to the server, which encapsulates the processed results again and returns them to the client. (3)Data persistence layer The data persistence layer is a relatively independent area for data persistence. It not only manages the mapping from the Java language to the database, but also provides a way to query teaching data and data acquisition for young children, to greatly reduce the time for manual development using JDBC and SQL to process data. This layer is mainly responsible for data interaction with the underlying database system, including the logical processing of system data. Accept the web server’s request, complete the database query and modification operations via the internal interface, and transfer all processed results to the server. The operation of the system is inseparable from the network architecture, which provides a blueprint for the design, construction and management of a communication network. In the design of the network framework, the visual teaching system of children’s basketball combines the actual needs of the construction of the basic basketball technology teaching course with the actual situation of children, at the same time, it can show a very strong security and convenience, so that users can carry out visual learning in the school’s external network or internal LAN at any time according to the actual situation of children. The network architecture of the system is shown in Fig. 2. As shown in the figure, the network setting of the visualized teaching system for toddler basketball can be operated either within the local area network or outside the campus network department. It is consistent with the characteristics of basic basketball teaching techniques, and also meets the current environmental conditions for teaching basic toddler basketball techniques. In order to meet the normal operation of the system in different network environments, the system comprehensively designs the network architecture from the campus LAN and the external network. From the perspective of the internal architecture, the server is divided into a database server and a WEB server, and internal communication circulates information through a network switch. Outside the local area network, if you want to implement the teaching of children’s basketball courses, you need to enter the teaching system through measures and settings such as firewalls and VPNs, which include network devices such as switches and routers. In this data circulation process, all follow the HTTP network protocol.

296

Y. Qin et al.

Teacher user

Web server Student Centralusers switch

Student users

Database server

Outside kindergarten

Basketball lessons

In kindergarten

Teacher user

router Central switch Firewall

Student users

Fig. 2. Network architecture of visual teaching system

3 The Hardware Design of Visual Teaching System for Children Basketball Course In the visual teaching system of children’s basketball course, the most important one is the visual interaction module, whose function realization is mainly based on the human– computer interface. Let children in the use of the software to learn basketball, full of fun, to achieve human interaction. In the design, the human-computer interaction module adopts the intelligent display terminal dmt80480507 OSW, which integrates the LCD, touch panel, data processing unit, storage unit, communication unit, etc. it can realize the human-computer interaction function quickly and conveniently. Dmt80480507 OSW user interface is shown in Table 1. Its control mode adopts full duplex asynchronous serial port, namely RS-232C interface. Table 1. Interface pin settings Pin

Function

Pin

Function

VCC

Power supply

DINT

Serial input

GND

Ground

DOUT

Serial output

BUSY

Buffer full

RS-232C is a serial data transmission bus standard for communication between computers and modems. It is one of the standard interface devices for computers. In the RS-232C standard, the signal uses a negative logic level, that is, a 3 V–-15 V is a logic 1, and a + 3 V–+15 V is a logic 0. DMT804805070 OSW implements the standard RS232C regulations. The logic 1 level is -5 V–-7 V, the logic 0 level is +SV–+7 V, and the serial interface of STM32F103ZET6 uses TTL logic level, so the system needs to realize and intelligent display through RS-232 transceiver Terminal serial communication.

Design of Data-Driven Visualization Teaching System

297

MAX3232 is a low-voltage differential full duplex RS-232 transceiver. In the range of +3 V–+5 V power supply voltage, only four 0.1 uF external connections are needed to form a dual voltage pump, reaching the highest communication rate of 120 Kbps, fully meeting EIA/tia-232 standard. The pin is as shown in Fig. 3, where C1+, C1, C2+, C2- are capacitor pumps, and the capacitance of the voltage pump is usually 0.1 uF.

C1+ Vs+ C1C2+ C2VsT2out R2IN

1 2 3 4 5 6 7 8

16 15 14 13 12 11

Vcc GND T1out R1IN R1out T1IN

10 9

T2IN R2out

Fig. 3. MAX3232 pin package

Use MAX3232 to realize RS-232-based data communication as the human-computer interaction interface of the embedded system, as shown in Fig. 4.

C1 C2 Uart1Tx Uart1Rx Uart4Tx Uart4Rx

1 3 4 5 11 12 10 9

C1+ C1C2+ C2T1IN R1out T2IN R2out

Vcc GND V+ VT1out R1IN T2out R2IN

16 15 2 6 14 13 7 8

GND

+3.3V C3 C4 PC-Rx PC-Tx HMI-Rx HMI-Tx

Fig. 4. RS-232 drive circuit

On the processor side, the data receiving and transmitting ends Uart4Tx and Uart4Rx of UART4 are connected to T21N and R20UT of MAX3232 and controlled by the system for serial communication, while T20UT, R2IN of MAX3232 and DMT804805070 OSW’s RS-232 transceiver HMI-RX, HMI-TX is connected to realize RS-232 communication with intelligent display terminal. In addition, because the MAX3232 is a dual RS-232 transceiver, another RS-232 transceiver is used as a debugging interface in the design to complete the data communication between the system and the computer, and to realize the functions of online programming and data storage of the system.

298

Y. Qin et al.

4 The Software Design of Visual Teaching System for Children Basketball Course 4.1 Visual Data Compression for Basketball Lessons Teachers can observe children’s learning of basketball technology through the visual module, and the basketball knowledge in the visual teaching system is mainly displayed in front of teachers and children through video, text and model. In the system operation, the above data needs to occupy a lot of system resources. The data compression algorithm is used to compress the data, reduce the useless use of resources when the system is in use, and improve the utilization rate of system resources. Let the compressed data size be Qbitand divide it into n blocks for compression. The size of each block of data is Qu , the proportion of Qbit data is Ru , and the block compression rate is αu . The compressed data size is: n 

Qu ·αu = Q ·

u=1

n 

Ru ·αu

(1)

u=1

Set the read-memory SRAM throughput rate as VSRAM , the decompression module throughput rate as Vdec (input terminal processing data speed), and configure the interface throughput rate as Vit . Then the actual throughput rate is Vc , and the above units are bit/s. Then the relationship between the above throughput rates can be expressed as: ε1 =

Vit VSRAM

(2)

Vit Vdec

(3)

ε2 =

In the formula: ε1 and ε2 are the relationship coefficients of throughput. Suppose the compression transmission size used by the system is Qbit, the time required for data is Tc , and the acceleration ratio of the data compression configuration is κ. The speedup ratio is expressed as:  Q Vit ·Tc , Vit ≤ VSRAM Tc κ = (4) Q VSRAM ·Tc , Vit > VSRAM Data compression function is added in the software design of the system, and appropriate acceleration ratio and other relevant parameters are configured to achieve the purpose of optimal data compression. 4.2 Data-Driven Data Consistency Processing Data driven technology can drive the change of basketball course knowledge through data and keep the change of basketball course knowledge consistent with the change of data.

Design of Data-Driven Visualization Teaching System

299

Using a data-driven selection collection, Selection is a collection of query filters from a tree-like scene structure, and provides a series of meta operations to manage this collection. Data joins can bind data to the knowledge of children’s basketball courses, allowing data joins to generate corresponding processing sets. These processing sets are important for changes in the structure of the scene and changes in the binding elements. At the Selection level, it focuses on conversion rather than performance. Selection focuses on the binding between data and toddler basketball courses. Selection provides data operation, which can bind the input data to the selected nodes. The system software design provides standard agnostic processing methods. The user input data is an array with any value as the element. Once the input data is bound to the node element, the meta operations on other selections can be operated by predicate. These operations can set or get attributes, styles, properties, selectors and text content. Generally, by default, the input data and the selected node correspond one by one according to the input data index, and the first node corresponds to the first datum of the input data. When the input data node set needs more accurate mapping, the key method of meta operation can be used to specify the mapping relationship between the input data and the node set, and the matched one will maintain mapping transformation Consistency in the process. The above process ensures consistency during system data changes. So far, the design of a data-driven visualized teaching system for toddler basketball courses has been completed.

5 The Performance Test of Visual Teaching System for Children Basketball Course The visualized teaching system test of the infant basketball course adopts a comparative test method, and the system resource consumption level is used as a test standard. The traditional infant basketball course teaching system is referenced and tested under the same test conditions. 5.1 Testing Environment The system test environment is as follows, the server system is ubuntul5.04, the compiler uses gcc4.9.2, the debugger uses gdb7.9, the browser uses Firefox 37.0, Lenovo desktop computer, Intel Core i5-3470 processor, 4 GB memory, 3.2 GHz frequency. The specific environment is shown in the following figure (Fig. 5). Before testing the system performance in the above test environment, the system integration test shall be carried out first. 5.2 System Integration Test The system integration test mainly tests the coordination and integration between the various modules of the infant basketball visual teaching system and the stability of the overall operation of the system to ensure that the system can run normally and stably

300

Y. Qin et al.

Fig. 5. System test environment

in the performance test. By using the system’s ability to actually access the data, the system’s blocking ability and pressure can be judged. The access time of children’s basketball teaching system to network data is shown in the table below (Table 2). Table 2. System access data time Times Time Whether the block was successful 1

0.02

Yes

2

0.07

Yes

3

0.13

Yes

4

0.21

Yes

5

0.04

Yes

6

0.06

Yes

7

0.09

Yes

8

0.14

Yes

9

0.17

Yes

10

0.15

Yes

From the data in the table, it can be seen that the access to the network data of the early childhood basketball teaching system can be blocked in time to ensure the safety of the network data. At the same time, it also shows that the designed visual teaching system of the early childhood basketball course based on the data-driven has good integration and coordination among all modules, and runs stably and normally to meet the needs of subsequent tests.

Design of Data-Driven Visualization Teaching System

301

5.3 Performance Testing The system performance test is mainly the system resource consumption test, which simulates the normal use of users and tests whether the system resource consumption is abnormal. The third-party software is used to monitor the resource consumption level of different systems in use. The traditional web-based teaching system test is used to record the results as test result 1. The traditional JSP based teaching system test is used to record the results as test result 2. The designed data-driven visual teaching system for children’s basketball course test is used to record the results as test The result is 3 (Fig. 6). 1.0

System resource consumption rate

0.9

1

0.8 0.7 0.6

2

0.5 0.4 0.3 0.2

3

0.1 0

20

40

60 Time/s

80

100

Fig. 6. Performance test results of different teaching systems

Observe the resource consumption of different teaching systems in the figure. Test result 1 shows that with the change of system operating time, its system resource consumption changes significantly. The initial time resource consumption is the largest, reaching 0.9, after reaching a peak, it gradually declines, and after it drops to 0.25, it starts to rise again. Within the time range, most of them are in a state of high resource consumption. Test result 2 shows that the system resource consumption in the first 20 s and 80–100 s of the system operating time changes greatly, and it is in a significant rising stage in the first 20 s, at 20–80 s, although there are some fluctuations, they are basically stable, ranging from 0.5–0.65. Between 80–100 s, there is a significant downward trend; test result 3 shows that during the system operating time, system resource consumption is not Large changes, the fluctuation range is between 0.2–0.3. Combined with the above data analysis, it can be seen that the system resource consumption level of the visual teaching system for children’s basketball course designed is significantly lower than that of the traditional teaching system, which shows that the visual teaching system for children’s basketball course designed based on data-driven is better than the traditional teaching system.

302

Y. Qin et al.

6 Concluding Remarks The kindergarten ball sports course focuses on exploring and forming a special course for ball games for all children, and optimizing the content and methods of teaching and learning in the development of ball games to improve teachers’ actual education and teaching level. To promote the overall healthy growth of young children. Designing a data-driven visualization system for children’s basketball courses provides great help for children’s basketball education and learning. Fund Projects. The construction and development of the open online course of “Children’s Basketball” under the background of vocational education. Project number: GXZZJG2019B112. Director: Yunfei Qin.

References 1. Jinyin, C., Zhen, W., Jinyu, C., et al.: Design and research on intelligent teaching system based on deep learning. Comput. Sci., 46(S1), 550–554 + 576 (2019) 2. Xiaofang, Y., Jing, Z.: Review of Chinese Martial Arts Education Research in the 40 years of reform and opening-up—visual analysis based on CNKI. J. Guangzhou Sport Univ. 39(03), 73–77 (2019) 3. Jin, H., Ning, Z.: A study on the relationship among preschool teachers’ mathematical teaching beliefs, pedagogical content knowledge and teaching behaviors. Teach. Educ. Res.31(05), 16–22+15 (2019) 4. Kunqi, J., Zhihua, W., Shuai, F., et al.: Data-Driven architecture design and application of power grid cyber physical system. Pow. Syst. Technol. 42(10), 3116–3127 (2018) 5. Jianfen, P.: Multi-objective unit teaching design and implementation-Taking forward handchanging basketball dribbling for example. J. Phys. Educ. 26(03), 115–120 (2019) 6. Jian, K. Yanglinxiao, H.: Research on evaluation and optimizing counter measures of University Network teaching information ecosystem. Mod. Inform., 39(07), 27–36 + 43 (2019) 7. Nan, L., Shiyun, Z.: Teaching and training platform for virtual equipment of replenishment based on Web3D. J. Syst. Simul. 31(06), 1136–1141 (2019) 8. Huifang, X.: The influence of deep learning on children’s use of scientific learning methods in collective and regional activities. Educ. Sci. 35(02), 72–77 (2019) 9. Daohua, L., Yushuang, C., Yansong, Z., et al.: Method for retrieving the teaching image based on the improved convolutional neural network. J. Xidian University (Nat. Sci.) 46(03), 52–58 (2019) 10. Xuejing, G., Kaifeng, Z., Yucheng, G., et al.: Design and implementation of ASD children’s cognitive education system based on unity3D. J. Syst. Simul. 31(05), 893–900 (2019)

Design of Distance Multimedia Physical Education Teaching Platform Based on Artificial Intelligence Technology Lu-zhen Mo1 , Yun-fei Qin2(B) , and Zhu-zhu Li1 1 Nanning University, Nanning 530200, China 2 Department Sports and Education, Guangxi College of Physical Education,

Nanning 530012, China [email protected]

Abstract. The traditional neural network-based education platform has the problem of slow access to resources, which affects teaching efficiency. To this end, artificial intelligence technology is introduced for the design of distance multimedia sports teaching platforms. Through system hardware and software design, enhance the integration capability of the teaching platform, improve the controllability of the teaching platform, help the physical education platform to obtain more educational resources, accelerate the processing and integration analysis of educational information, and better assist teachers in developing physical education jobs. In order to verify the effectiveness of the method in this paper, the method in this paper is compared with the teaching platform based on neural network,the experimental results show that the platform can obtain resources quickly and has broad application prospects. Keywords: Artificial intelligence · Remote multimedia · Physical education

1 Introduction At present, the pace of development of society is getting faster and faster, so saving people’s time and space, and facilitating people’s learning and living has also become the design purpose of scientific and technological researchers. In order to save the school’s education space and optimize the teaching force, a curriculum education platform using Internet technology came into being [1]. In the rapid development of the Internet era, the application of new technologies such as big data and cloud computing is becoming more and more widespread, which not only brings huge changes to people’s lives, work, and entertainment, but also has a profound impact on the development of China’s education [2]. In the context of the era of big data artificial intelligence, the continuous emergence of Internet technology and learning software has led to the rapid development of online education models, and education informatization has become an inevitable trend of the future development of education in China [3]. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved S. Liu et al. (Eds.): eLEOT 2020, LNICST 340, pp. 303–314, 2020. https://doi.org/10.1007/978-3-030-63955-6_26

304

L. Mo et al.

The original online distance education was based on the postal, radio and television. With the development of multimedia technology and computer technology, the current distance online education is a novel educational model based on Internet technology. As for long-distance physical education, domestic scholars and experts have done little research [4]. However, due to the need to provide a large venue for sports activities, and the classification and borrowing of sports equipment is cumbersome and complicated, the human resources of the school have not been used more reasonably. With the innovative research and development of new technology education products, various education industries are actively constructing education platforms, innovating teaching models, and using traditional education and big data methods to form new teaching models [5]. The platform uses big data storage, processing and processing technology to provide rich teaching resources for school education, and has changed the problems of lack of traditional teaching resources, poor teaching quality, and low learning efficiency. The platform’s teaching model has injected new Vitality is of great significance to the development of education [6]. How to build a long-distance multimedia physical education teaching platform under the background of big data, increase physical education teaching resources and teaching management, improve physical education teachers’ teaching quality and student learning efficiency, and create a professional and personalized intelligent learning environment is an urgent problem for the education industry. In view of these developments, this paper designs and develops a long-distance multimedia sports teaching platform based on artificial intelligence technology.

2 Construction of a Distance Multimedia Sports Teaching Platform 2.1 Overall Platform Architecture The long-distance multimedia sports teaching platform based on artificial intelligence technology is completed based on big data, artificial intelligence technology and mobile Internet technology. It not only has a large number of rich physical education teaching resources, but also provides students with personalized learning services [7]. Students can operate at any time and place and find physical education teaching content and teaching videos according to their learning needs. This vivid, flexible and convenient learning method effectively stimulates students’ interest in learning. Moreover, students can also communicate with physical education teachers and classmates through this platform, increase the interactivity and autonomy of physical education activities, and effectively stimulate students’ enthusiasm for autonomous learning [8]. The hardware structure of the long-distance multimedia sports teaching platform includes servers, storage, user terminals, network equipment, and so on. The software functions include sports education information deployment and database development. The platform can realize the sharing of educational resources, fully reflects the comprehensive and personalized service system, and lays the foundation for users to provide efficient services. For the construction of physical education teaching platform, a lot of technical support is needed to realize the sharing of resources, data, and information. The application of artificial intelligence technology not only provides rich learning resources, but also saves students’ time. It also plays a long-distance multimedia physical education teaching platform. High-quality performance [9].

Design of Distance Multimedia Physical Education Teaching Platform

305

The overall architecture design of the platform is shown in Fig. 1.

Fig. 1. Overall platform architecture

(1) Physical layer The physical layer is the core level of the hardware structure of the long-distance multimedia sports teaching platform. The hardware includes servers, servers, and network equipment. Through the mining, storage, analysis, and processing of sports education resource data, the correctness and security of rich information resources are guaranteed [10]. (2) Virtual resource layer This layer is adjacent to the physical layer. The main hardware facilities include storage resource pools, computer resource pools, data resource pools, and network resource pools. They are responsible for the preservation and operation of platform information resources. (3) Logical layer This level is responsible for platform system task management and resource management. It is the core management of the remote multimedia sports teaching platform. It is also responsible for system fault detection and repair to ensure the platform’s safe and stable operation. (4) Presentation layer The presentation layer is an intermediate layer. The platform module includes a presentation module, an access module, and a visualization module. It provides services for mobile clients and web portals through three modules. The access module is mainly responsible for authorizing access and configuration, the visualization module is responsible for converting online confidence into data form and performing visual processing, and the display module is to display the visual processing results to the user, and is responsible for personal information management, registration and login, personalization Settings, column content display and management, etc., and responsible for pushing information resources and content search.

306

L. Mo et al.

(5) Application layer The application layer also belongs to the middle layer and is the core level of the entire platform. It mainly includes physical education, multimedia learning, intelligent management, intelligent services, and multimedia environments. It provides professional teaching resources for education and services for teacher-student communication. This level contains a wealth of sports information resources, which can meet the full service of sports education and sports learning. (6) Network layer A network interface is set inside the network layer to provide users with a main path to access the cloud platform. (7) User layer This level is at the top of the hierarchy, and users get relevant information through mobile terminals. 2.2 Hardware Structure Design The long-distance multimedia physical education teaching platform is based on the Internet of Things hierarchical structure system. First, the overall design of the physical education teaching platform is adopted. The artificial intelligence technology is used to collect the original information of the physical education teaching platform. The information processing module realizes the serial bus control of the physical education platform. The hardware structure of the physical education platform is designed in the Multigen Creator 3.2 development environment. The network module of the physical education teaching platform is based on the IoT architecture. The campus network and local area network are used as the network layer of the smart classroom platform. The Microsoft Visual Studio development component is used to develop the TCP/IP protocol for the

Fig. 2. Hardware structure

Design of Distance Multimedia Physical Education Teaching Platform

307

physical education teaching platform. The functional modules of the designed physical education teaching platform include power supply module, personnel attendance module, LED display system, asset management system, etc. The TCP/IP protocol and UDP protocol are used for the network design of the physical education teaching platform, and the central centralized controller is used to realize the communication between the physical education platform and the computer network. Based on the above analysis, the hardware structure is shown in Fig. 2. According to the hardware structure of Fig. 2, the remote multimedia sports teaching remote automatic control system adopts the local bus control method. The system is modularly designed in the embedded environment, and then the hardware development of the system is performed. The remote multimedia sports is combined with the embedded ARM processor. Integrated design of teaching and computer control. The hardware modules of remote multimedia physical education include sensor module, integrated control module, bus module, interface module and power module. Mesh network architecture method is used for input and output control and integrated processing of various devices in remote multimedia physical education.

3 Platform Development The introduction of artificial intelligence technology in the distance education platform provides a more vivid teaching environment for teachers and students, and enriches the education model. The distance physical education teaching platform must first understand the characteristics of physical education teaching flexibility, and design the specific teaching method and content of the course based on such teaching characteristics, as well as the design of curriculum assessment. Curriculum design requirements can fully mobilize students ‘enthusiasm for learning, and can guide students’ courses, especially the standardized guidance for action classes is clear and thorough. 3.1 Development Environment Description The multimedia sports teaching platform adopts the intelligent design scheme of remote automatic control and bus integrated control. The designed sports teaching platform is built on the Multigen Creator 3.2 and embedded ARM development environment. It uses artificial intelligence technology to process the intelligent information of the sports teaching platform And integrated information analysis. The multimedia teaching information is transmitted through the integrated information processing terminal for bus transmission and information integration, and intelligent communication design is performed on the host computer module. The technical indicators of the designed sports teaching platform are as follows: (1) Establish a VIX bus module for integrated scheduling and transmission of control instructions. The main frequency of the sports teaching platform is 24 MHz, the design power of the platform is 5 kW, and the power magnification is 80 dB. (2) The sampling rate of the bus control is greater than 200 kHz. The D/A resolution for multimedia control of the sports teaching platform is not less than 12 bits.

308

L. Mo et al.

(3) Design the RFID time and attendance machine. The resolution of RFID for data collection is about 13 dB. It has the function of remote attendance and tag identification. The output static power loss is 20 W. (4) The UHF RFID card reader is designed for error control, and the output error level is 12 dB