<![CDATA[Research on junior-to-senior career transitions in badminton remains limited, particularly in quantifying conversion rates and applying machine learning for predicting transition success. This study analyzes medal records of 147 Indonesian players across four junior tournaments (BWF World Junior Championships, Asian Junior Championships, Suhandinata Cup, Youth Olympic Games) and two senior tournaments (BWF World Championships, Olympic Games) spanning 1992–2024. Descriptive analysis reveals an overall junior-to-senior conversion rate of 29.3%, with significant variation across medal count groups (χ² = 36.84, p < 0.001). Players with four or more junior medals achieve 51.7% conversion rate compared to 0% for single-medal winners. Random Forest and XGBoost classifiers, trained with SMOTE oversampling and evaluated via repeated stratified cross-validation (5-fold, 10 repeats), achieve 90.8% and 88.8% accuracy respectively, substantially outperforming a logistic regression baseline (83.6%). Junior gold medal count and BWF World Junior participation emerge as the strongest predictors. A sensitivity analysis excluding recent junior medalists confirms the robustness of these findings. These results provide empirical evidence for optimizing talent identification in Indonesian badminton development programs. Keyword: Badminton, Career Transition, Conversion Rate, Random Forest, Talent Identification, XGBoost]]>
Copyrights © 2026
