Based on K-Means Clustering and Time Series Datadriven Approach Construction of a Future Award Prediction Model

This study looks at the various factors that influence Olympic medal performance, using statistical models to provide insights and recommendations for the future. We initiated the development of the national medal count model by performing a comprehensive data preprocessing. After the initial data cleaning, we started feature engineering to optimize the construction of the random forest regression prediction model. Given the extensive nature of the original dataset, which contains information about the Olympic Games from 1896 to the present, we only identified and retained data from 1960 to 2024 to ensure relevance and temporal validity. Subsequently, we aggregated the number of participants per year and the total number of MEDALS awarded by each country, while binary encoding participation in specific sporting events. In order to improve the prediction accuracy of the model, the K-means clustering algorithm is further applied to the original medal data to classify countries into high, medium and low competition levels. Finally, the improved method successfully constructed the random forest regression model to predict the medal table of the 2028 Olympic Games. This paper points out several noteworthy phenomena, including the distribution pattern of gold, silver and bronze, and the significant positive correlation between the competitive level of a country and the number of participants.