Ethical AI in Sports Analytics: Medal Projection and Coaching Effects via SHAP Regression

This document proposes a machine learning-based Olympic medal prediction model using feature engineering and the Random Forest method, achieving over 95% accuracy. SHAP analysis quantifies key factors, revealing the influence of event numbers and host advantages. The Gradient Boosting Tree model further explores historical trends and the “great coach effect,” demonstrating the significant impact of experienced coaches on medal performance. These findings provide strategic insights for Olympic Committees and national teams.