Model to Predict the Result of a Soccer Match Based on the Number of Goals Scored by a Single Team

Alba Maribel Sánchez Gálvez, Ricardo Álvarez González, Sully Sánchez Gálvez, Mario Anzures García


Soccer is a very popular sport; it is a fine subject of study given the large amount of data it generates. This article presents a model that through Machine Learning algorithms predicts the victory or defeat of a soccer team, based on the number of goals scored. This model applies four machine learning classifiers: Linear Regression, Support Vector Machines, Naive Bayes and Decision Trees. The proposal is supported with data from the Mexican football league from 2012 to March 2020, the study has been divided into two sections: in the first draws are considered and in the second aren’t, with the purpose of discovering the influence of draw in analysis. With the proposal model accuracy in the range of 81% to 84% was achieved without draws and considering ties the accuracy was in the range of 72% to 75%.


Supervised learning, machine learning algorithms, assessment metric

Full Text: PDF