Presenter(s)
Pronob Kumar
Abstract
Predicting soccer match outcomes is difficult due to the low-scoring nature, frequent draw outcomes, and other factors. This project used a historical dataset from European leagues spanning 1960 to 2024 to develop supervised machine learning models for predicting match results. The dataset included various features related to match conditions, team information, and historical performance. The study compared multiple machine learning models, including logistic regression, decision trees, XGBoost, and random forest, with a regression-based baseline model. Model performance was evaluated using accuracy, confusion matrices, ROC curves, and AUC metrics. Preliminary results indicate that machine learning models outperform the baseline model in predicting match outcomes, particularly for home and away wins, while draw outcomes remain more difficult to predict.
College
College of Science & Engineering
Department
Computer Science
Campus
Winona
First Advisor/Mentor
Trung Nguyen
Second Advisor/Mentor
Mingrui Zhang
Location
Kryzsko Great River Ballroom, Winona, Minnesota; United States
Start Date
4-23-2026 2:00 PM
End Date
4-23-2026 3:00 PM
Presentation Type
Poster Session
Format of Presentation or Performance
In-Person
Session
2b=2pm-3pm
Poster Number
32
Prediction of Soccer Match Result with Machine Learning
Kryzsko Great River Ballroom, Winona, Minnesota; United States
Predicting soccer match outcomes is difficult due to the low-scoring nature, frequent draw outcomes, and other factors. This project used a historical dataset from European leagues spanning 1960 to 2024 to develop supervised machine learning models for predicting match results. The dataset included various features related to match conditions, team information, and historical performance. The study compared multiple machine learning models, including logistic regression, decision trees, XGBoost, and random forest, with a regression-based baseline model. Model performance was evaluated using accuracy, confusion matrices, ROC curves, and AUC metrics. Preliminary results indicate that machine learning models outperform the baseline model in predicting match outcomes, particularly for home and away wins, while draw outcomes remain more difficult to predict.
