Research on Student Sports Training Data Modeling and Personalized Training Program Optimization Based on Multi-Layer Perceptro

Abstract

In this paper, based on the Gray Wolf Optimization (GWO) algorithm, Cauchy variation operator and cosine convergence factor are added, and the convergence speed of the algorithm is enhanced by the position updating formula to shorten the training time, and the improved multilayer perceptron is used for modeling student sports training data. After that, from the perspective of user groups, user-based collaborative filtering algorithm (UB-CF) is selected to model the sportsmen. Then from the perspective of sports, CB recommendation algorithm is used to build recommendation object model based on sports features, and finally UB-CF algorithm and CB algorithm are combined to form a personalized sports recommendation algorithm, so as to achieve the purpose of personalized recommendation to users. The results show that when the overlap rate is 75% and the window size is 3/4 of the original data length, the classification accuracy of the x-y dataset can reach 97.32%. The recommendation effect in the personalized training program of student sports shows that the value of RMSE of this paper's recommendation algorithm (0.1016) is much lower than that of the comparison method, and its predicted value is highly consistent with the actual value. It can be seen that the method in this paper makes the improved recommendation algorithm more perfect and the recommendation accuracy higher by deeply mining the user's preference.