Medal prediction model based on machine learning and Bayes

The Olympic Games, organized by the International Olympic Committee, is the largest summer comprehensive games in the world, and its medal list has attracted much attention. The Olympic Games is a dynamic and complex system, and it is of extensive and far-reaching practical significance to establish a scientific and accurate prediction model for the competition results and to reveal the rules of medals. In this regard, this paper will address the following issues. For Problem 1, we first used Machine learning algorithms and Random Forest models. The goodness-of-fit index was used to judge the advantages and disadvantages of Random Forest, Logistic regression and XGBoost, and secondly, we predicted the number of medals won by each country and the number of medals won by each country in 2028, and with the help of the correlation analysis and the systematic clustering algorithm, we came up with the intrinsic connection between the host country, the amount of project changes and the amount of medal changes. For problem 2, we firstly adopt Bayesian Changepoint Detection monitoring model. We use Bayesian Changepoint Detection monitoring to determine the location of the effect point of "great coaches", then we use the factor of coach's contribution rate to determine the influence of coaches in national programs, and at the end of the question, we have conducted case studies on China, England and Brazil, and verified the reasonableness of the model by combining with the real situation in history. For question 3, we first summarized the model above, provided insights related to the Olympic medal count, and explained how each type of insight informs the Olympics. The host country's home field effect and international economic power were analyzed, and we thus made recommendations to the Olympics on infrastructure development, logistical experience, and so on, in order to provide for the next Olympic Games in Los Angeles, USA.