Soft Computing-Based Result Prediction of Football Games

Abstract

Soft computing methods for result prediction of football games based on fuzzy rules, neural networks and genetic programming techniques, are proposed in the article. The models are taking into account the following features of football teams: difference of infirmity factors; difference of dynamics profile; difference of ranks; host factor; personal score of the teams. Testing shows that the proposed models achieve a satisfactory estimation of the actual game results. The current work concludes with the recommendation of soft-computing techniques as a powerful approach, either for the creation of result prediction models of diverse sport championships, or as effective data extrapolation mechanisms in case of limited available statistics.