Football analysis using machine learning algorithms
Loading...
Date
2025-10-09
Authors
Χύτας, Χρήστος
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In recent years, technology has been growing rapidly. Surely all of us have heard about Artificial Intelligence (AI). In recent years AI has also entered the field of football. Many innovations that help the game have been implemented like Goal Line Tecnhology, Semi-Automated Offside, and Video Assistant Referee (VAR). Of course, technology is being used from the teams as well, to improve their performance and preparation. A significant role in this can be played by Machine Learning, which is a subset of Artificial Intelligence. Machine learning is a branch of artificial intelligence that gives computers the ability to learn from data and improve through experience, without being explicitly programmed. Since medical data most of the times are private, it was very difficult to find a suitable dataset that could fulfill our needs. So we decided to create a custom dataset, from injury history data that were extracted from transfermarkt, FIFA player stats, the English Premier League 2023-2024 schedule, as well as some weather condition data. In this thesis, we will try to figure out how easy is it to predict a possible upcoming injury, by using few machine learning algorithms Specifically, after using the SMOTENC technique to deal with undersampling, since our dataset is highly imbalanced, we are implementing the following 4 algorithms: KNN, Random Forests, Gradient Boosting and a Neural Network. All 4 algorithms provided results close to the ones that other authors achieved, and taking into consideration the high level of difficulty for a prediction like this one, can be considered good. Among them, KNN and Neural Network provided the best performance, with Random Forests following, and Gradient Boosting at the end, having the worst performance.
Description
Keywords
Machine learning, Artificial intelligence, Football, Algorithms, Injury prediction, K-Nearest neighbors, Random forests, Decision trees, Gradient boosting, Neural networks