This study aims to investigate the main features of European male and female soccer focusing on match actions data. A framework for objective and unbiased feature extraction and comparison is presented focusing on explainable algorithms. Match technical attributes by player were collected from event data and categorized by game period and player position. Supervised learning methods were used to induce the differences between male and female data points and the results were obtained using machine learning interpretability methods to understand the underlying mechanics of the models implemented. This study focuses on presenting a novel framework for comparison interpreting the black box of machine learning. The comparison results aim to provide an initial data-driven descrip-tion of the difference between male and female tactical behavior in European soccer as well as a baseline for cross-country tactical soccer comparison. The results could serve for better and customized design of trainings and match strategies.
© Copyright 2020 spinfortec2020digital: 13. Symposium der dvs-Sektion "Sportinformatik und Sporttechnologie" vom 24.-25. September 2020 in Bayreuth. Veröffentlicht von Institut für Sportwissenschaft. Alle Rechte vorbehalten.