A study of momentum prediction based on particle swarm optimisation and random forests

This paper proposes a PSO-RF based momentum strategy prediction model. Firstly, the model quantifies the performance of players, determines the key factors affecting the players’ scores through principal component analysis, constructs the momentum quantification model, and conducts data processing and visual analysis. Then, PSO is introduced to optimize the RF model to solve the overfitting problem, feature labels are extracted and analyzed for prediction, and strategic recommendations are given. Finally, through the extension analysis and result analysis of the model, the accuracy and generalization ability of the model are verified.