Application of Decision Trees in Athlete Selection: A Cart Algorithm Approach

Abstract

This study investigates the application of Decision Trees (DTs), a non-parametric supervised learning method, renowned for its simplicity, interpretability, and wide applicability in various domains, including machine learning for classification and regression tasks. The focus of this study is on the use of DTs, employing the Classification and Regression Trees (CART) algorithm, in the initial screening of athletes. This involves analyzing 11 sociodemographic and anthropometric variables within a dataset of 113 prospective athletes, encompassing both numerical and categorical data. The DT model exhibits outstanding performance, achieving accuracy and precision rates exceeding 0.8. Further analysis, varying impurity criteria and tree depths, indicates that the Gini index at a depth of 3 optimizes accuracy. Notably, weight, and Body Mass Index (BMI) exhibit the highest significance among the other variables. Looking ahead, future research could explore enhancing DTs' predictive capabilities in athlete selection by incorporating more variables or employing ensemble learning techniques. This study lays the groundwork for further investigations aiming to refine athlete screening processes and broaden the utility of DTs in sports-related predictive modeling.