Ervin Vladi\u0107, D\u017eelila Mehanovi\u0107 and Elma Avdi\u0107<\/p>\n
Abstract
Predictive modeling in football has emerged as a valuable tool for
enhancing decision-making in sports management. This study
applies machine learning techniques to predict football match
outcomes in the WWIN League of Bosnia and Herzegovina. The
aim is to evaluate the effectiveness of various models, including
Support Vector Machines (SVM), Logistic Regression, Random
Forest, Gradient Boosting, and k-Nearest Neighbors (kNN), in
accurately predicting match results based on key features such
as shots on target, possession percentage, and home\/away status.
By (1) gathering and analyzing match data from three seasons, (2)
comparing the performance of machine learning models, and (3)
drawing conclusions on key performance factors, we demonstrate
that SVM achieves the highest accuracy at 83%, outperforming
other models. These insights contribute to football management,
allowing for data-driven strategic planning and performance
optimization. Future research will integrate additional factors
such as player injuries and weather conditions to improve the
predictive models further.<\/p>\n
\u00a0<\/p>\n\n\n\n