Analysis of Player Performance and Game Fluctuation Based on LM-BP Neural Network and SVM

In this study, the 2023 Wimbledon Men's Singles Final was used as the object of study, and the intense match between Carlos Alcaraz and Novak Djokovic was comprehensively analyzed by using LM-BP neural network and SVM model. In the data preprocessing stage, classification, cleaning and correlation analyses were used to ensure the accuracy of the data. In terms of player performance assessment, the LM-BP neural network was used to quantify the players' serving and returning abilities, and momentum was used as a quantitative indicator of how good or bad a player's performance was. The results show that momentum is highly similar to the trend of player scoring rate and has a significant correlation with the difference of match player scores (correlation coefficient of 0.772, significance test p<0.001), which proves the key role of momentum in the match. For the analysis of match fluctuation, the SVM model was used to predict match fluctuation and the MRMR algorithm was used to determine the importance of three variables, namely, unforced errors, the number of breaks of serve and the number of aces, on match fluctuation, which was successfully predicted in the case of the final match with an accuracy of 83.673%.