A Machine Learning Based Momentum Prediction Model for Tennis Matches

In tennis, "Momentum" is an important research field. Through the establishment of an evaluation index system and the use of logistic regression to analyze the significance of the relationship between the index system and the scoring results, machine learning models are utilized for 5-fold cross-validation to develop a model that can capture the emergence of points during the game. This is used to determine the strengths, weaknesses, and performance levels of players on the field, providing visualizations to describe the game's course based on the model. Explore metrics that can help determine when a match trend shifts from favoring one player to another and utilize match data to construct a model that predicts match trends.