Predicting Football Outcomes
with Bayesian Networks
M. van Dijk (TU Delft - Electrical Engineering, Mathematics and Computer Science)
Gabriela F. Nane – Mentor (TU Delft - Applied Probability)
E.M. van Elderen – Graduation committee member (TU Delft - Mathematical Physics)
Leo Iersel – Graduation committee member (TU Delft - Discrete Mathematics and Optimization)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
In this thesis Bayesian Networks are used to predict European football matches between the years 2008 and 2016. The goal of this research is to see how the structures learned by different Bayesian Network learning algorithms influences the predictions. First the data is explored and modified to be used for Bayesian Networks and secondly the theory is explained using examples. Finally the theory is applied on the data and the structures are learned with the help of a bootstrap method and the predictions are validated using 5-fold cross validation. We can conclude that the networks learned by the algorithms and with the help of an expert give a good representation of the underlying relationships, but are not very good in prediction the end result.