AN EVALUATIVE MODEL BASED ON ARTIFICIAL NEURAL NETWORKS TO ASSESS MOTOR COORDINATION IN ROTATION PHASE OF YOUNG DISCUS THROWERS

Abstract

In recent years, there has been a growing interest in using artificial intelligence (AI) techniques to evaluate athletic performance. Artificial neural networks (ANNs) enable the detection of complex movement patterns that are difficult to measure using traditional tests . This research aims to build an evaluative model based on ANNs to assess motor coordination during the rotation phase in young discus throwers and to study its relationship to performance . The researcher employed a descriptive-analytical approach with a sample of 30 young discus throwers. Multilayered neural networks were used to train the model on a set of kinematic variables extracted from video recordings, and the network outputs were compared with expert estimates. The results indicated that the model possessed high scientific coefficients and revealed a strong correlation between the extracted motor coordination scores and numerical achievement (r = 0.88). Six standard levels ranging from "weak" to "excellent" were also constructed and can be used in evaluation and selection processes. The study recommends using this model to evaluate young discus throwers and improve training programs.