Application of Artificial Intelligence in Soccer Tactical Analysis and Player Performance Evaluation

Abstract

Tactical analysis and player performance evaluation in soccer are undergoing a shift from traditional experiencebased judgment to a combination of data and theoretical support. This paper systematically reviews and categorizes two mainstream technical pathways and one novel technical pathway in the current field of football analysis: video and computer vision analysis methods, big data analytical methods, and the ecological dynamics method. Exploring the underlying principles, structural mechanisms, and respective weaknesses and strengths of the various technical methods, this paper points out that diverse methods have their own differences in research objectives as well as application contexts and show a high degree of complementation. Moreover, this paper suggests that these ways can be implemented in various areas, such as the assessment of player performance, tactics, and the work with the team in the long-term. Finally, the paper presents the problems of football analytics in terms of data quality, model interpretability, integration of methods, and the development of talent based on interdisciplinary talent. It also describes the future trends of the convergence of various technologies.