Use of Machine Learning Algorithms to Predict the Results of Soccer Matches

The prediction of football results is a very complex phenomenon to model. Historically, models based on regression or classification of data have been proposed. However, these proposals present limitations in terms of the number of data that can be used, the type of factors that need to be considered in the models, and the intrinsic complexity of a model that is by nature random. An alternative to these proposals is to use models based on machine learning techniques, which have the advantages of being able to process a larger number of data and the ability to learn patterns of behavior from the data. In this article, we present a web application that aims to provide predictions about the results of football matches using different machine learning techniques. The application also provides a set of additional services that allow users to browse and visualize information about matches, teams, and players; functions to manage the application's information about matches, teams, and players; and a game for registered users based on predicting the results of upcoming matches. When predictions are correct, users earn points that allow them to be ranked in a user ranking.