基于知识图谱的体育人工智能研究分析

Analysis of Researches on the Application of Artificial Intelligence Technology in Sports Field Basedon Knowledge Graph

以Web of Science数据库1994年以来“体育”“锻炼”“运动”“机器学习”“深度学习”“计算机视觉”等关键词为主题的926篇文献为数据来源,利用“Cite Space V”软件进行可视化处理和分析,以知识图谱的方式梳理近25年的体育人工智能研究,探讨体育人工智能研究的进展和发展方向。认为:1)体育人工智能研究地区分布较广,美国处于世界领先水平,中国的研究质量有待提高。2)体育人工智能研究的高产作者与团队集中在美国高校,以开发与完善针对不同人群的,基于机器学习与深度学习技术的智能穿戴设备为主要研究方向。3)体育人工智能研究涉及到多个学科,主要运用和借鉴工程学、计算机科学和体育科学的研究方法和理论。4)体育人工智能研究的热点分为三大聚类群,具体是体质健康促进、运动损伤防控和运动能力提升。研究载体主要以基于IMU的智能穿戴设备和基于GPU的计算机视觉分析为主。研究算法从机器学习算法逐渐转向深度学习算法。

With the data source of 926 articles themed with keywords such as“physical education”,“exercise”,“sport”,“machine learning”,“deep learning”and“computer vision”in the Web of Science database since 1994,through visual processing and analysis with Cite Space V software,this paper analyzes the researches on the application of artificial intelligence technology in sports field in the past 25 years by means of knowledge graph,and discusses the progress and development direction of sports artificial intelligence research.The conclusions are as follow:1)Sports artificial intelligence research areas are widely distributed,with the United States in the world’s leading level,and China’s research quality in need of improvement.2)The prolific authors and teams of sports artificial intelligence research are mainly from universities in the US,with the development and improvement of intelligent wearable devices based on machine learning and deep learning technologies for different populations as their main research direction.3)The research of sports artificial intelligence involves many disciplines,mainly using and drawing lessons from the research methods and theoretical perspectives of engineering,computer science and sports science.4)The research hotspots of sports artificial intelligence are divided into three clustering groups,including“physical health promotion”,“prevention and control of sports injuries”and“improvement of sports ability”.The research carriers are IMU-based smart wearable devices and GPU-based computer vision analysis.The research algorithm gradually transforms from machine learning algorithm to deep learning algorithm.