An Exploration of Momentum Changes in Tennis Players Based on CART Decision Regression Trees

Abstract

As the sport of tennis is booming, the study of players' technical and tactical level and game status is becoming more and more important. For this reason, for the sake of the research of imposing changes. First of all, from the six tennis technical and tactical evaluation indexes of scoring change, error change, serve-receive scoring change, technical change, momentum change, and streak change, this paper derives the four characteristic indexes that are most relevant to the target variables, i.e., serve-receive scoring change, technical change, scoring change and streak change, through the CART algorithm. Among them, the correlation between the change in serve-receive score and the change in momentum was the greatest. Then, a decision regression tree model based on the CART algorithm was established. Then the appropriate decision tree model parameters were obtained after 10 rounds of data validation by the 10-fold cross-validation method. Among them, the depth is 5 and the minimum number of leaf nodes is 20. Meanwhile, to prevent the model from overfitting, this paper carries out a pre-pruning operation on the decision tree model. That is, estimation is performed before each node division, and if the division of the current node cannot bring the decision tree generalization performance improvement, the division is stopped and the current node is labeled as a leaf node to improve the applicability of the model. Finally, through the mean square error (MSE) and coefficient of determination (R-squared) test, it is found that the mean square error (MSE) is very close to 0, and the coefficient of determination R-squared is very close to 1, which indicates that the model has good prediction effect.