Momentum Study Based on LightGBM Model and Spearman Correlation Analysis

Momentum often influences the trend of the game, thus prompting us to think about how momentum affects the outcome of the game and the various factors that influence the game. Based on the above issues, this paper first categorized the influencing factors into two types: positive and negative influencing factors and created a momentum assessment model that used the entropy weighting method to determine the influence weights of each factor and plot the momentum of each participant over time. Then, a running test and Spearman correlation analysis were conducted to verify the association between momentum changes and game outcomes. In addition, this paper defines turning points, implements fluctuation prediction, and verifies the accuracy of turning points through gradient boosting decision tree-based algorithms, logistic regression, random forest, and support vector machine models. Thus, it provides training suggestions for players to reduce the occurrence of unforced errors for future matches. Finally, the accuracy of the model in this paper was evaluated and the model was applied to other matches, having verified the predictive ability of the model.