Abstract:

The aim of this study is to predict the ranking of the Turkish Super League team with the artificial nervous networks (YSA) model developed according to 7 different input variables. In the study, a total of 918 league matches played in three seasons (2015/2016, 2016/2017, 2017/2018) in the Turkish Super League; the number of passes (suitable passes, attack passes and goal-pre-passes), the holding of the ball throughout the match, the time of attack resulting in the goal and the number of shots thrown were assessed. The study analyzed the matches played in the 2015/2016 and 2016/2017 seasons (input variables) and the 2017/2018 season league ranking (output variable) was predicted. The value produced in the model is in the range of 0 - 1 for a trained network has been found a league ranking by 100 for a trained network. According to the analysis results made with the developed YSA model, the Turkish Super League team ranking was estimated with a accuracy rate of over 94% for many teams in the test data set. As a result, the most important factors that change the score in football have also been found to be an important parameter in determining the league ranking of the shut, the winning ball, the number of passes, the time of attack and the winning ball. Thanks to the match analysis with the YSA model, it is believed that coaches can make significant outcomes based on performance.

Keywords: