Abstract:

Purpose: The study aims to calculate the percentage of the variables affecting the match outcome by data mining methods and artificial nerve networks of the male and female volleyball super League matches played in Turkey and Italy between 2013 and 2020. Tool and Method: This research is correlational scanning research. 1144 of them are Turkish men, 1142 of them are Turkish women, 1122 of them are Italian men, 1066 of them are Italian women, with a total of 4474 competition data mining program in Matlab, the "Statistics and Machine Learning Toolbox", the "Neural Network Pattern Recognition Toolbox" vehicle packages and the "Classification Learner" application being analyzed. Classification methods are applied differently for all matches, gender and country, and classification accuracy is determined with the "20 fold cross validation" technique with 10 different variables affecting each match within itself. The findings: The research found that the rate of correct prediction of the match results with the techniques used in the results was 90%. The most influential variables in the match result are the variations in the game of the moving teams. The result: The results found that the match between data mining programs and artificial nerve network modeling resulted in the correct predictions of approximately 90%. When we look at the men's matches, the most influential variables in the match result are the blocks of the shipping teams, their attacks and the number of services. When we look at women’s matches, the two most important factors affecting the outcome of the matches are the movement team attacks and the movement team service numbers.

Keywords: