Research on Olympic Medal Distribution Prediction Based on Cluster Regression and CUSUM Analysis Model

It is well-known that the Olympics, as a significant international sporting event, attracts widespread attention to the medal tally. Therefore, predicting Olympic medal counts is of utmost importance. Consequently, this paper constructs three models to analyze and forecast the trends in medal rankings, the “great coach” effect, and gain original insights into the Olympics. Firstly, data preprocessing is carried out, followed by identifying key features to predict medal counts. For countries that have never won medals, we face a “cold start” problem. To address this, clustering algorithms are applied to identify similar situations, and then models are trained to learn the transition pattern from 0 to 1. By combining clustering algorithms with the trained models, methods such as linear regression, random forest regression, support vector regression, and Lasso regression are employed for data fitting, and the performance of different models is evaluated. To explore and quantify the significant role that “great coaches” play in medal outcomes, this paper analyzes the performances of the Chinese and American volleyball teams, as well as the Romanian and American gymnastics teams. Through quantitative analysis, we obtain values for the influence effect and conclude that great coaches can significantly improve a country's performance in specific events, thereby increasing the chances of winning medals. Furthermore, we apply the CUSUM method to analyze fluctuations in medal counts. Using examples from different events in China, the United States, and France, we provide recommendations. Based on the results of the model analysis presented in this paper, insights into the trends in Olympic medal counts are proposed.