Author(s)

Lingxin Kong, Chuan Ju, Diming Wu

Corresponding Author

Lingxin Kong

Abstract

This paper investigates the prediction of tennis match momentum transitions based on multiple linear weighting and logistic regression. To analyze the match situation more comprehensively, we considered the scenarios of consecutive scoring and losing points and introduced 11 new variables to measure match momentum. Through Principal Component Analysis (PCA), a dimensionality reduction technique, we consolidated 14 variables into 9 key components to simplify the model and enhance its generalizability and conciseness. Furthermore, we employed the Analytic Hierarchy Process (AHP) to assign weights to independent variables, which were then used for stepwise regression analysis to select influencing factors. By applying multiple linear weighting, we constructed a performance scoring model and predicted match turning points using a logistic regression model. We defined "swings in the match" as changes in the state of a particular player and built a model to quantify the impact of various indicators on match swings and predict the specific moments when the win rate transitions from one state to another. Our experimental results demonstrate that the proposed model has high accuracy in real-time win rate prediction and momentum transition prediction. We conclude that the model can provide effective predictions for tennis match momentum transitions, offering valuable insights for coaches and athletes.

Keywords

Tennis Match Analysis; Momentum Transition Prediction; Multiple Linear Weighting; Logistic Regression; Principal Component Analysis; Analytic Hierarchy Process