Meezan: An AI-Assisted Coaching System for Greco-Roman Wrestling Using a Norm-Referenced, Factor-Based Elite-Distance Framework

Résumé

Artificial intelligence (AI) has transformed numerous domains of sport science, with recent advances in language and vision-language models opening new avenues for athlete assessment and coaching support.Greco-Roman wrestling, an individual, one-on-one combat sport that demands both physical fitness and technical skill, is particularly well suited to AI-driven, individualized training and assessment; yet the literature on AI-enabled coaching systems for this sport remains sparse.In practice, a coach's central need is a means of gauging how far a given wrestler's current profile diverges from an elite reference standard, across both fitness and skill dimensions.This paper presents Meezan -a system that addresses this need.We assembled a dataset of 91 elite Greco-Roman wrestlers, aged 13-17, from wrestling clubs in Iraq, each assessed on a validated 31-item physical and technical test battery.Exploratory factor analysis reduced this battery to twelve latent factors, independently reproducing loadings reported in the source instrument.We then developed a norm-referenced, factor-based elite-distance framework, using a robustcovariance Mahalanobis distance to quantify how atypical a wrestler's overall profile is relative to the elite reference group, with an exact per-factor decomposition identifying which factors drive that distance.This framework was deployed in a web-based coaching system that computes individualized and team-level assessments and generates natural-language coaching reports via a large language model.We view this work as one step toward more robust, AI-assisted coaching systems, to be extended with larger and more diverse datasets and additional coaching objectives 1 .

Citer ce document

Accès au document

Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter

Auteur(s)

Statistiques

Consultations : 2

Téléchargements : 0