Machine learning application in soccer: a systematic review

Abstract

Due to the chaotic nature of soccer, the predictive statistical models have become in a current challenge to decision-making based on scientific evidence. The aim of the present study was to systematically identify original studies that applied machine learning (ML) to soccer data, highlighting current possibilities in ML and future applications. A systematic review of PubMed, SPORTDiscus, and FECYT (Web of Sciences, CCC, DIIDW, KJD, MEDLINE, RSCI, and SCIELO) was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. From the 145 studies initially identified, 32 were fully reviewed, and their outcome measures were extracted and analyzed. In summary, all articles were clustered into three groups: injury (n = 7); performance (n = 21), which was classified in match/league outcomes forecasting, physical/physiological forecasting, and technical/tactical forecasting; and the last group was about talent forecasting (n = 5). The development of technology, and subsequently the large amount of data available, has become ML in an important strategy to help team staff members in decision-making predicting dose-response relationship reducing the chaotic nature of this team sport. However, since ML models depend upon the amount of dataset, further studies should analyze the amount of data input needed make to a relevant predictive attempt which makes accurate predicting available.

Exploring the interplay between metabolic power and equivalent distance in training games and official matches in soccer: a machine learning approach

“…Of course, the manipulation of the area per player (ApP), the presence of the goalkeeper, or the design-specific rules contribute to increase or decrease the position-specific demands concerning the desired external load outcomes ( Castellano and Casamichana, 2013 ; Lacome et al, 2018 ; Riboli et al, 2020 ). However, this study’s novel approach of using ML analysis to distinguish between TGs and official matches based on MP and EDI provides new insights into the differences between these two types of soccer activities (TGs and official matches, Jaspers et al, 2018 ; Rico-González et al, 2023 ). Furthermore, the significant correlations between Edwards’ TL and the TL based on MP metrics support the validity of using both parameters to measure ETL in soccer ( Table 3 ).…”

“…Thirdly, the study did not investigate the relationship between external load and performance outcomes, which is essential for understanding the practical implications of the findings. Finally, the main limitation of this study is the use of convenience sampling, aware that ML models depend upon the amount of dataset ( Rico-González et al, 2023 ); consequently, this investigation should be regarded as a case study.…”

“…Over the past decade, ML has emerged as a popular approach in sports science research for analyzing complex data sets in sports performance analysis. ML is a powerful tool that can help sports scientists and coaches gain valuable insights into player performance and make data-driven decisions to enhance athletic performance ( Rico-González et al, 2023 ). Therefore, the main goal of this study is: 1) to examine the interplay between MP and ED and their relative contribution to TGs and official soccer matches; 2) the secondary aim of this study was to examine the relationship between the external training load (ETL), determined by the interplay between MP and ED, and the internal training load (ITL) estimated using the HR-based method, which was used as a criterion validity measure.…”

SoccerNet: A Gated Recurrent Unit-based model to predict soccer match winners

“…Their proposed model achieved an accuracy of 46.6% in the test dataset. In 2023, Rico-González et al carried out a systematic review on the application of ML models in football [ 29 ]. They focused on the use of ML algorithms in injury prediction, match winner prediction, and talent hunt prediction.…”

Injury Prediction in Sports using Artificial Intelligence Applications: A Brief Review

“…As innovative schemes are accomplished with applying novel rules more efficiently, it is thus necessary to regularly gather and examine literature that has been used for injury prediction and prevention or that may be applied to these purposes in the future. In addition, even though recent literature reviews have been exploring specific facets of this industry, there are still some limitations: the majority of the papers are printed from the viewpoint of data mining [5], they are game-oriented [7][8][9], they have a partial possibility [3,4,10], or they only concentrate on group game [6]. Our goal is to provide an exhaustive review of the current status of machine learning (ML) in sports injury research, spanning various sports and utilizing various algorithms.…”