Résumé:
This thesis aims to design an intelligent model based on artificial intelligence and
machine learning techniques to guide football players towards optimal playing
positions based on their physical and morphological characteristics. The study
adopted a descriptive survey methodology, analyzing data from 938 players in the
French League (Ligue 1 and Ligue 2) for the 2022-2023 sports season. Physical and
morphological measurements were collected, and machine learning algorithms
(KNN, SVM, Random Forest) were applied to classify players into positions based
on physical metrics. The results revealed that the machine learning models were able
to classify players with an accuracy ranging between 70.21% and 89.36%, with clear
physical differences observed between various positions. Central defenders were
characterized by height, strength, and jumping ability; midfielders excelled in
endurance and agility; while attacking players were distinguished by speed,
acceleration, and explosive power. The results also showed a slight superiority of the
Random Forest algorithm in most classification scenarios. The study provides a
practical framework that coaches and football academies can rely on to enhance
sports selection and orientation processes
