The Estimation of the Set Results in 2016/2017 Vestel Venus Sultans League Games by Artificial Neural Network

Abstract

The objective of this study is to estimate the set result via Artificial Neural Network (ANN) by considering the scores of the volleyball teams at technical time-outs (8th and 16th points) and 21st point. In the study, 132 games, 984 sets and 4152 points that were played and scored during a season by 12 teams playing in 2016/2017 Vestel Venus Sultans League were examined separately. 85% of all sets that teams played in one season were randomly reserved for training and 15% for test. Verbally winning or losing was modeled as 0 (zero) or 1 (one) numerically. Since the produced value was between the ranges of 0 – 1, for a trained network, it was multiplied with 100 and thus the possibility of winning was obtained. Consequently, it was determined that the developed model estimated the set result for many teams (test dataset) with an accuracy rate over 95%. By means of competition analysis to be made using ANN model in volleyball, it is thought that technical officers can reach fast and accurate conclusions at the moment of the set is played. It can be said that these conclusions will provide technical officers with a warning mechanism to take necessary technical and tactical measures while the set is being played.