Medal Prediction Analysis Based on Random Forest and ARIMA Models

In this study, various algorithmic models such as Random Forest and ARIMA are combined, and the study aims to establish an accurate medal prediction model. Firstly, the data of past competitions were preprocessed, multiple features and predictor variables were selected, and categorical features were coded. Then, using Random Forest, LightGBM and XGBoost models, we fit the data through five-fold cross-validation, construct the relationship between features and predictor variables, and then integrate the prediction results of the three models using the stacking model to successfully construct the Olympic medal prediction model. At the same time, with the help of ARIMA and gray prediction model, we predicted the feature variables of the competition and substituted them into the medal prediction model to get the medals of each country. This research method considers various factors, and the model has high prediction accuracy, which provides a new and effective idea for Olympic medal prediction.