Research on Momentum Evaluation and Change Point Prediction in Tennis Matches Based on PCA and EWM-Topsis Model

This paper introduces a model for momentum evaluation and change point prediction in tennis matches based on Principal Component Analysis (PCA) and the Entropy Weight Method combined with the Technique for Order of Preference by Similarity to Ideal Solution (EWM-TOPSIS). The tennis match dataset is preprocessed through a series of steps, including the transformation of categorical variables into continuous variables, imputation of missing values, and detection and treatment of outliers, ensuring the integrity of the dataset and the accuracy of subsequent analyses. The features of the momentum model are then extracted, and key performance indicators of players during the match are effectively identified through PCA dimensionality reduction. The EWM-TOPSIS method is applied for a comprehensive evaluation of player momentum, determining the weights of serving, stamina, and hitting as primary factors. Additionally, Bayesian change point detection technology is utilized to successfully predict turning points in the match, and the Random Forest algorithm is employed to analyze key features affecting player scoring. The study concludes that serving quality, the current game score difference, and momentum values are significant factors influencing player scoring, providing a new perspective and decision support for tennis match analysis.