A Hybrid Machine Learning Approach for the Modeling and Prediction of the UEFA EURO2020

Conventional   approaches   that   analyze   and   predict   the results of international matches in football are mostly based on the framework of Generalized Linear Models. The most frequently used type of regression models in the literature is the Poisson model. It has been shown that the predictive performance of such models can be improved by combining them with different regularization methods such as penalization.

More recently, also methods from the machine learning field such as boosting and  random forests turned   out   to   be   very   powerful   in   the   prediction   football   match outcomes. Here, we analyze both a hybrid random forest extension based on conditional inference trees and a hybrid boosting extension based on  extreme gradient boosting  for modeling football matches. The models are fitted to match data from previous UEFA  European Championships  (EUROs) and based on the corresponding estimates all match outcomes of the EURO 2020 are repeatedly simulated (100,000 times), resulting in winning probabilities for all participating national teams.