Predicting Olympic medal counts with data-driven modeling

The Olympic Games, a grand quadrennial sports events, always attract people's attention. As a sports fan, Olympic medal table intrigues me a lot. Therefore, this study builds a comprehensive evaluation model based on multidimensional data, incorporating historical medal counts, performance fluctuations, and national sports potential assessments to predict the medal table of the next Olympic Games. This article creatively proposes a method for grouping countries based on their Olympic participation history and medal winning trends, establishing a national classification model and a potential evaluation model. After that, using a Random Forest regression model to predict medal counts for the 2028 Los Angeles Olympics, incorporating adjustments for the host country effect. Furthermore, the paper introduces a Shapley value-based approach to quantify the influence of different sporting events on medal distribution. Finally, sensitivity analysis is conducted to assess the robustness of predictions under varying external conditions. These findings provide valuable insights into Olympic medal prediction and strategic planning for national sports development.