A Momentum Decision Recommendation Model based on Random Forest and Game Theory Methods

Abstract

Momentum, as an important indicator of athlete status, has been widely used in sports competitions. This paper chose tennis games to establish standard momentum measuring model, momentum forecast model, tennis game model based on mixed strategy simultaneous decision game under momentum influence. First, this paper constructed a standardized physiological and psychological measurement model for player momentum using critic weight method and coefficient of variation method, which is highly correlated with score. Second, this paper used the random forest algorithm to predict the trend of the competition and obtained the weights of the factors that affect the momentum. Lastly, this paper further introduced the concept of game theory and established a payment matrix based on mixed strategy simultaneous decision game to calculate the Nash equilibrium solution of both parties under the influence of the current momentum. According to the models, this paper can predict the trend of the athlete's status and prompt the coach or the athlete to make a more correct choice according to the status of both sides.