Athlete performance prediction using intelligent reporting and regression model analysis: A generative approach for training planning.

Abstract

This study explores the use of reported data and predictive analysis as a long-term generative approach for athlete training planning. The data collected from 607 higher education students (mean age = 16.86; Std = 1.22), includes measurements from physical tests and activity records. The dataset comprises 29 variables, which were profiled and feature-engineered to enhance predictive accuracy for training procedures. We leveraged Microsoft Azure Machine Learning to determine the significance of features on outcomes and utilized Power BI to visualize the impact of aggregate features on running distance. Initial findings indicates that the optimal age range for focusing training efforts is between 16 and 17 years old. This result is supported by a Spearman correlation coefficient of 0.42, stipulating a moderate positive relationship between the age group and the predicted performance outcomes based on key aggregate features. In particular, four key features significantly impact performance, while other variables have a minor influence. The study highlights the significance of these aggregate features in predicting training success. In conclusion, the study underscores the importance of a robust reporting process and the use of predictive analysis in developing training programs. It identifies four critical features that have a substantial impact on realized performance. While these four features are paramount, the study also acknowledges that other variables, though less influential, may still affect the outcomes if considered. This comprehensive approach to data collection and analysis provides a solid foundation for optimizing athlete training programs, ensuring that training efforts are both targeted and effective. The findings offer valuable insights for coaches and sports scientists aiming to enhance athletic performance through data-driven training strategies.