Performance Prediction of Speed Skaters Based on BP Neural Network

Abstract

To better predict the performance of skaters, a BP neural network-based approach is proposed. With the help of the main component factor analysis method, the performance sample of the athletes was reduced, the traditional high-dimensional data set was changed to the new low-dimensional data set, the size of the model input variables was reduced, and the analysis efficiency was increased and improved. Experimental results show that the larger the iteration step, the lower the prediction error, because more a priori information of sports performance is used through multiple iterations, which improves the prediction accuracy. The error of sports performance prediction using this model is less than that of traditional methods, and the maximum difference is 35.98%. It is proved that this method uses factor analysis and neural network training to cluster and fuse the data information of sports performance, improves the prediction accuracy, and has a good performance in convergence and robustness.