Using neural networks and hierarchical cluster analysis to study goal kicks in football

Abstract

The objective of this study is to build a mathematical model that predicts the success of a goal kick in soccer. The model is based on an ensemble of neural networks whose inputs are five features extracted directly from the goal kick and one more that depends on the opposing team. This new variable is calculated using a hierarchical cluster analysis and divides the possible opponents by their defensive strategy. It is shown that it has a relevant importance in the result of a goal kick, which validates the idea of using the presented methodology that takes into account the opponent’s tactics when analyzing specific plays of a soccer match.