Tennis Player Momentum Prediction Based on Logistic Regression and Decision Tree Models

Tennis is a fast-paced sport, and in this paper, momentum calculation is accomplished and validated by developing a model that can quantify momentum and visualize real-time momentum in tennis. Firstly, the grey correlation between the indicator coefficients and the reference volume is calculated and analyzed, and the momentum indicator with the greatest impact on the score is selected, so as to visualize the momentum value and assess the real-time performance of the player on the sports field. In order to verify a certain correlation between momentum and score, a mixture of metrics was selected to derive a correlation between momentum and score using statistical analysis. Finally, in order to predict the problem of fluctuation and change in game performance, logistic regression model and decision tree are constructed for score prediction.