Medal Prediction Analysis Based on Random Forest and DID Models

This paper focuses on medal table prediction and related strategy analysis, aiming to provide decision-making advice to National Olympic Committees of various countries through data analysis. Algorithmic models such as multiple linear regression, random forest regression, and individual fixed-effects model are used in the study. First, principal component analysis was used to reduce the dimensionality of the data indicators, and the multiple linear regression and random forest regression models were optimized by grid search to predict the medal standings, and the prediction results of the random forest model were recognized based on RMSE. At the same time, the model was established to predict the probability of change in the number of medals of each country and the probability of winning for countries that did not win. Secondly, using Chow test and difference-in-differences model (DID), the validation and estimation of relevant effects are carried out. The strength of this study lies in the comprehensive use of multiple models to analyze medal-related situations from multiple perspectives, which provides a comprehensive and in-depth perspective for the study of sports events.