Research on the Impact of Momentum on Game Situations Based on Random Forest and XGBoost Models

Abstract

The momentum in the realm of sports is like an intangible force, unleashed by a sequence of events. During a game, a team or player may feel they are riding this wave as if victory is within reach. However, the formation of momentum and its impact on the outcome of the game remains a mystery. To study the factors that influence momentum and explore their impact on it, ultimately understanding how they affect the course of a game, This study establishes a Random Forest Model to calculate the weights of these independent variables which we use to define the momentum and XGboost model to detect momentum combined with using SHAP Values to analyze the impact of different feature quantities on match fluctuations to learn about the impact of variable parameter changes on results. Finally, we improve strategic decision-making through genetic algorithm optimization based on the XGboost model to find the optimal allocation of indicators that maximize player momentum. This research has the ability to help us figure out the significance of momentum exploration in competitive scenarios and how to optimize momentum.