Analyzing tennis match momentum through machine learning techniques and the TOPSIS model

Momentum plays a crucial role in tennis performance, yet its quantification and prediction remain challenging in sports analytics. This study develops a comprehensive framework to analyze momentum patterns in the 2023 Wimbledon Men's Singles tournament through an integrated approach combining random forest machine learning and TOPSIS methods. The research identifies and evaluates key momentum indicators encompassing player performance metrics, score dynamics, and physical exertion parameters. Our random forest model processes these features to generate initial predictions, which are then refined using the TOPSIS method with entropy-based weighting. The hybrid model, tested across multiple datasets, demonstrates exceptional predictive accuracy with R-squared values of 0.892 for training and 0.845 for testing. Following initial validation, the framework was enhanced by incorporating environmental variables such as court surface characteristics and weather conditions, improving its practical relevance. The findings demonstrate that momentum is not only quantifiable but serves as a significant predictor of match outcomes, offering valuable implications for strategic planning and performance optimization in professional tennis.