Research on Athlete Momentum Prediction Based on CNN-LSTM Model

This paper focuses on evaluating the scoring rate of players in a tennis match using a neural network model combining CNN and LSTM. First, outliers and missing values were processed, and the data were normalized to improve the model accuracy. Second, a momentum scoring system was constructed based on covariance analysis, and a CNN-LSTM model was used to predict the performance of players at a specific time. On this basis, the model was evaluated by ROC curves and correlation analysis to demonstrate the effect of momentum on the match results, and the generalization ability of the model was explored to provide a research basis for the generalization of the model to other matches in the future. Finally, the correlation between momentum and the scoring rate of each ball was analyzed by using the Pearson correlation test and the SHAP method was used to interpret the predictions of the model.