Implementation of K-Means Particle Swarm Optimization for Clustering Football Players in the Top Five European Football Leagues

Football Clubs are prioritize victory and championship aspirations, requiring skilled players proficient in goal-scoring, precise passing, error reduction, and tactical comprehension. The objective of this study is to identify and analyze the most effective clustering approach and performance outcomes of football players using the combined methodology of k-means clustering and particle swarm optimization. The research data employed in this study consists of spatio-temporal information pertaining to the performance of football players. The outcomes of the k-means clustering, in conjunction with PSO, employing values of k ranging from 2 to 4, reveal that cluster k \(=\) 3 represents the most desirable group of players for recruitment recommendations by football clubs, with a silhouette coefficient of 0.76. Furthermore, the clustering process yields Cluster 1, comprising players capable of playing in three positions, Cluster 2 consisting of players specialized in one position, and Cluster 3 encompassing players adept at playing in two positions.