Medal Prediction for the Olympic Games Based on Grey Prediction Model and Markov Chain

Abstract

This study aims to predict the medal counts for the Olympic Games by employing a combination of the Grey Prediction Model and Markov Chain. The Grey Prediction Model is utilized to forecast the total medal counts for various countries, taking into account multiple covariates such as the number of participating athletes, the number of events participated in, the score of national athlete performance, and the top five sports specialization scores. These covariates are derived from historical Olympic data through meticulous data pre-processing and feature engineering steps, which include handling missing values, standardizing athlete names, and conducting Monte Carlo simulations to estimate participation uncertainties. The Markov Chain model is then applied to predict the likelihood of countries winning their first Olympic medals, based on historical medal outcomes and relevant covariates. The results of the study provide projected rankings for the total and gold medal tables in the 2028 Los Angeles Summer Olympics, as well as estimates for the number of first-time medal-winning countries. Additionally, the study examines the relationship between sports specialization and medal success by analyzing the coefficients of the Grey Prediction Model. The findings highlight the significance of strategic event selection and specialization in achieving Olympic success and offer valuable insights into the competitive landscape of the upcoming Games.