Momentum Prediction Model Combining Markov Chain And MLP

In sports competitions, it is common for a certain player to dominate and score consecutively, which can be described as the effect of "momentum". To portray the abstract momentum to describe the change of athletes' state during the competition, and thus provide technical support for the subsequent training of athletes, a volume prediction model combining Markov chain and MLP is proposed. Firstly, the resultant variables are calculated by the entropy value method for momentum scoring, after that, technical indicators are considered, Markov features are calculated by the Markov chain model and added into the MLP model for momentum prediction, and finally, the important factors determining the momentum are analyzed by SHAP method. The experimental results show that the accuracy of predicting athletes' state reaches 98%, which can provide credible race reviews and guidance suggestions.