Predicting football match outcomes with ensemble machine learning models
Authors
Patel, Dhiren
Issue Date
2024
Degree
MSc in Business Analytics
Publisher
Dublin Business School
Rights holder
Rights
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Abstract
The study performed here focused on the prediction of the results of football matches using machine learning models. The machine learning models considered in the study were Random Forest (RF), Decision Tree (DT) and the Gradient Boosting classifier and these models were used to build an ensemble model which was the model that performed prediction based on voting. The data associated with football matches was used for training the ensemble model. The best 20 features from the data were selected using the Chi-square technique and the class imbalance in the dataset was solved using Synthetic Minority Oversampling Technique (SMOTE). The results of the study showed that the ensemble model showed an accuracy of 99.5% in predicting the results of football matches. The model was implemented as a desktop application that predicted if the outcome of the football match was a win, lose or draw for the home team.