Research on Olympic Medals Prediction Model Based on Linear Regression and Logistic Regression

Abstract

This paper proposes a linear regression model, a logistic regression model, and a counting model with fixed and random effects for Olympic medal prediction, focusing on the application of different models in predicting the number of medals of each country in the 2028 Olympic Games, the possibility of winning a medal for a country that has not yet won a medal, and the influence of excellent coaches on the number of medals. First, a linear regression model was used to predict the number of gold, silver and bronze medals of the 2028 Olympic Games for 15 countries by estimating the regression coefficients through the least squares method with the number of historical medals and the number of athletes as the characteristic variables. Secondly, logistic regression model is applied to extract the characteristics such as the number of athletes and the number of events, and a binary classification model is built through maximum likelihood estimation to predict the probability of winning for the countries that have not won any medals, and the 10 countries that are most likely to win gold medals are given. Finally, a counting model containing fixed effects and random effects is constructed, and the parameters are estimated with the help of Bayesian method to quantify the influence of excellent coaches on the number of medals. The model system can accurately predict Olympic medals from different dimensions, and the combination of multiple algorithms effectively improves the accuracy and comprehensiveness of the prediction, providing scientific model support for Olympic medals prediction and related analysis.