A Study of Tennis Match Momentum Based on Regression Modelling and Random Forests

In this study, we use a coarse- and fine-grained regression model and a random forest algorithm in conjunction with a data pipeline to quantify momentum in tennis matches and assess its impact on match outcomes. We construct models that accurately capture the flow of the match, demonstrating the importance of momentum in the match. By analyzing the correlation of fluctuations in different factors and predicting subsequent fluctuations in the data, we reveal the decisive role of momentum and suggest targeted training and tactical adjustments. The application of these methods provides insights for understanding and optimizing athlete performance, highlighting the non-random impact of momentum on game outcomes.