人工智能技术在国际体育运动领域的聚类与演化

Clustering and evolution of artificial intelligence technology in the field of international sports

以Web of Science数据库1995年以来“体育人工智能”等关键词为主题的549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,以可视化知识图谱的方式梳理了近25年体育人工智能研究的国家、学科分布、研究热点以及演化趋势,探讨其研究进展和发展方向。1)体育人工智能研究地区分布较广,其中美国、中国和德国处于领先地位。2)体育人工智能研究涉及到多个学科,主要运用和借鉴了计算机科学、工程学、体育科学等学科的研究方法和理论视角。3)关键词的频次与中心性印证了目前体育人工智能领域是以机器学习为主要方向,人工神经网络为主要算法,注重以数据挖掘为基础的实践与实证研究。4)研究热点包括基于可穿戴加速度计技术的简单活动识别与能量消耗研究;基于可穿戴式传感器的动作分析与损伤防控研究;基于卷积神经网络算法的计算机视觉场景分类研究;基于计算机视觉的体能与技战术的分析与预测;基于计算机深度学习的人体姿态识别技术。

Taking 549 articles from Web of Science database since 1995 with the theme of"sportsartificial intelligence"as data sources,using Cite Space V software visualization processing and analysis,the authors sorted out the artificial intelligence research in the past 25 years in the form of visual knowledge map,including the countries that carry out the research,the subject distribution,the research hotspots and evolution trend,discussed the research progress and development direction.It is concluded that:1)Sports artificial intelligence research is widely distributed,among which the United States,China and Germany are in the leading position;2)The research involves many disciplines,mainly using and drawing on the research methods and theoretical perspectives of computer science,engineering,and sports science;3)The frequency and centrality of keywordsconfirm that the current field of sports artificial intelligence focuses on machine learning,artificial neural network as the main algorithm,and focuses on practical and empirical research based on data mining.4)Research focuses include research on simple activity identification and energy consumption based on wearable accelerometer technology;movement analysisand damage prevention and control research based on wearable sensors;classification of computer vision scenes based on convolutional neural network algorithm;analysis and prediction of physical ability,technique and tactics basedon computer vision;body posture recognition technology based on computer deep learning.