Olympic Competitiveness Assessment and Medal Prediction Based on TOPSIS Entropy Weight-Random Forest Model

With the continuous refinement of the Olympic system, existing medal prediction methods based on single indicators have become inadequate for assessing multidimensional competitiveness. Specifically, this study aims to establish a comprehensive evaluation system for evaluating medal-winning potential at the national level. Using the entropy method and PCA-TOPSIS approach, four core dimensions—ability, changes, competitiveness, and strategy—are extracted from twelve initial indicators to form a multidimensional evaluation index system. On this foundation, a Random Forest Regression model is employed to conduct a predictive analysis of Olympic medal distribution. The research findings indicate that the United States is expected to perform exceptionally well in both gold medals and total medals, maintaining its leading position. The model evaluation results, with a value of 0.9437 and an MSE of 0.0117, demonstrate a high degree of goodness-of-fit. Furthermore, in simulating predictions for the medal tally of the 2024 Olympic Games, the model's predictions deviate by less than 5% from actual data, further validating the model's effectiveness. The study reveals that this comprehensive evaluation system effectively overcomes the limitations of single indicators, providing a more scientific and comprehensive analytical approach to Olympic medal prediction. This research not only broadens the theoretical dimensions of sports competitiveness assessment but also offers a quantitative reference for countries to formulate Olympic strategies, thereby possessing significant theoretical and practical value.