Aplicação de técnicas de mineração de dados para análise de partidas de futebol

Abstract

The interest in predicting the outcome of soccer matches remains relevant today, especially given the growing sports betting market in Brazil and around the world. To determine favoritism in a match, a number of factors are considered, including player statistics, history of matches and performance during the season. However, the distance traveled by teams throughout the season for away games is often overlooked. In this study, we investigated the influence of the distance traveled by teams on the results of Premier League matches, using data mining techniques. The process involved creating a data warehouse with performance information and geographical data, supplemented with data from midweek competitions. Scenarios were defined and the logistic regression algorithm was applied to train the model. Analysis of the results showed that the distance traveled by the teams had no significant impact on match results. The comparison with bookmaker data showed similar accuracy, indicating that other factors may be more relevant in determining results. In conclusion, although the distance traveled by the teams is relevant to the logistics of competitions, it does not seem to be a determining factor in Premier League results.