Tan Qingwen. Football Player Performance Prediction Based on Combined Kernel Function Correlation Vector Machine. Dynamic Systems and Applications 29 (2020) No. 5, 2029 – 2040

https://doi.org/10.46719/dsa202029537

ABSTRACT.
The football game is two team-based each team consists of 11 players with the ball. In the football (soccer) team, the evaluation of players exchange the place, squad formation, and strategic team planning is very important. However, with wide pool level players at the grassroots level, short career pillars, across different conditions, playing different environments, venues, and various club budgets, it is difficult to determine the performance value of each player. The proposed Kernel Function Correlation Vector Machine (KFCVM) algorithm is designed to solve the above problems of player performance prediction.KFCVM extractsthe various attributes and abilities of football players based on previous matches each player performance. The proposed method helps coaches and team management identify future opportunities in football matches such as club budget, league competition, and the importance of players in that team or region without being biased by subjective conditions. The proposed system is useful for player performance in making assumptions about a football player dataset and the extract’s overall relationship between players’ attributes value, market value, and performance value. These ratings depend on the level at which the footballers are played and the skills they do possess based on the previous match analysis.

Keywords: KFCVM, Prediction, Football, Player performance, Attributes, Game