Document Type : Research Paper I Open Access I Released under CC BY-NC 4.0 license
Authors
1 Corresponding Author, Department of Psychology and Educational Sciences - Physical Education and Sports Science, Yazd University, Yazd, Iran.
2 . Department of Physical Education and sport science, National University of Skills (NUS), Tehran, Iran.
3 M.Sc., Department of Motor Behavior, Faculty of Physical Education and Sports Sciences, University of Tehran, Tehran, Iran
Abstract
Introduction: Sports talent identification (STI) is a complex process that benefits from insights across various scientific fields to enhance credibility. Nowadays, artificial intelligence has proven effective in prediction and classification tasks. Given that talent identification fundamentally revolves around accurately predicting and classifying individuals, leveraging AI can be truly transformative. This study aims to explore how artificial intelligence can be used for STI.
Methods: Despite the diversity of sports disciplines, four categories were selected: ball sports, racket sports, martial arts, and aquatic sports. The expert's opinion method was used to prioritize the selected indicators, which include anthropometric, somatic, motor control, biomechanics, physical and mental fitness factors. A total of 310 participants (boys and girls aged 6 to 16) were evaluated and labeled based on their scores. To assess the relationship between each component and the labels, independent t-tests were conducted, resulting in the selection of six key components for each category. Finally, the Perceptron 6-1-1 neural network was used to check the accuracy and validity of STI.
Results: The results of neural networks showed that the accuracy of the network for classifying people in ball sports, rocket sports, martial arts, water and other sport fields were 97.9, 97.9, 87.2, 91.5 and 80.8% respectively, which is a high and desirable accuracy.
Conclusion: Finally, it can be said that determining the principal components of each sports field and designing an artificial neural network helps researchers and coaches to know the important indicators of each field and use it for STI.
Keywords
- Artificial neural network
- Ball sports
- Martial arts
- Racket sports
- Sports talent identification
- Water sports
Main Subjects
)