Research on Momentum Dynamic Prediction Based on BP Neural Network and Support Vector Machine Algorithm

In this paper, the influence of player momentum on the result of tennis match is studied. Machine learning algorithm and mathematical calculation method are used to build momentum evaluation and prediction model for in-depth analysis. Firstly, BP neural network is used to fill in the missing data and establish an evaluation system to quantify the player's momentum. Secondly, a real-time momentum evaluation model is established by using support vector machine to help evaluate the dominance of both players. Finally, the random forest is used to predict the game situation, and the importance order of the factors’ influencing momentum is determined. It is found that the factors such as the server and forehand score have the greatest influence on momentum. Through the application of the Wimbledon women's singles data and the sensitivity analysis of the model, it is proved that the model has strong generalization ability and robustness, which provides guidance for athletes, coaches and data analysts.