METHODOLOGY OF TALENT SELECTION AND LONG-TERM PERFORMANCE FORECASTING IN BELT WRESTLING BASED ON ARTIFICIAL INTELLIGENCE ALGORITHMS

Abstract

The modern era of high-performance sports is characterized by a transition toward "Sports 4.0," where data-driven decision-making replaces traditional scouting. This paper presents a comprehensive methodology for talent identification and long-term performance forecasting in Belt Wrestling (Belbog‘li Kurash) using advanced Artificial Intelligence (AI) and Machine Learning (ML). Traditional selection methods often suffer from "coaching subjectivism" and a bias toward early-maturing athletes (Relative Age Effect). To solve this, we developed a multi-factorial diagnostic system based on a dataset of 450 wrestlers. The "Kurash-Intellekt" software, utilizing Artificial Neural Networks (ANN) and Random Forest algorithms, was designed to process 42 unique biometric and technical indicators. Our longitudinal study (2024–2026) demonstrates that the AI-driven model identifies elite potential with 87% accuracy, representing a 25.1% improvement over traditional pedagogical methods (p < 0.05). Furthermore, the integration of "Digital Twins" allows for the simulation of individualized adaptive training trajectories. The findings suggest that AI can effectively predict athlete success over a 7-year horizon by prioritizing neurodynamic stability and biological age over static physical strength.

Keywords

Artificial Intelligence, Machine Learning, Talent Selection, Belt Wrestling, Predictive Analytics, Neural Networks, Digitalization, Neurodynamics, Digital Twin, Sports 4.0, Morphofunctional Diagnostics.