A Novel AI-Driven Framework for Talent Identification in Cricket Batting Skills to Leverage Immersive Virtual Reality Performance Analytics

Abstract

The proposed study will utilize machine learning to create a model to predict the potential of cricket batsmen with the help of the combined and objective evaluation process. The conventional talent discovery approaches used in the field of cricket may be limited by weather changes, subjectivity, and poor evaluation on perceptual-cognitive skills. To limit these shortcomings, the current study suggests a hybrid model that would integrate technical skill assessment, field-based target-based testing of performance, immersive virtual reality (VR) simulations, and machine learning (ML) classification. The number of right-handed batsmen in the study was 50, at the Under-19 district level. Assessment of players was done through standardized technique test, Cricket Australia Centre of Excellence (CACOE) batting test and a customized VR-based perceptual-cognitive task. Thirty-two performance variables were obtained and they were employed to train a Random Forest classification model. Statistical analyses involved the comparison of the performance with different ball speeds, correlation analysis and ML-based categorization. Findings showed that a decrease of 23-28% in on-field performance was observed in an increased ball speed (110 km/h to 130 km/h; p < 0.001), with VR-based tests showing a higher level of consistency and accuracy on shots (81% vs. 57%). Random Forest model had a total classification accuracy of 87.4%. It is.important to note that VR speed of decision-making (r = 0.79) and dealing with short pitched high-speed deliveries (r = 0.58) stood out as the primary predictors of batting potential. The results indicate that VR combined with ML will make the objectivity, reliability, and scalability of the cricket talent identification systems more effective..