Tennis Match Prediction Based on Wavelet-BP Neural Network Model

In sports, assessing a player's performance during a match and the impact of momentum on the outcome of the match are key issues, but are often difficult to measure accurately. In this study, we first preprocessed the match data into serve and non-serve data and then developed a Wavelet-BP neural network model to analyze the player's performance at specific times of the match. It was finally obtained that Carlos Alcaraz performed best in the interval between [2500,5000] and [7000,8000] (s) of the match, and performed poorly after 10000 seconds. Meanwhile, by quantifying the momentum into indicators such as winning streak and whether to serve or not, and verifying its correlation with the match results by chi-square test, the results show that momentum has a significant effect on the match results. This study provides a new method to assess players' performance in a match and quantify the effect of momentum on the match outcome. This is important for improving tactical strategies in sports competitions, enhancing team competitiveness, and increasing the spectacle of the game.