Study On Time Prediction Of Chinese Elite Speed Skating Athletes In 1500 Meters By 30s Wingate Test

Research objective: This paper makes a comprehensive,systematic and comprehensive analysis of the athletes’ anaerobic exercise ability through the 30 s Wingate test,selects a number of sensitive indicators,and uses linear regression and neural network methods to predict the time spent of athletes in China’s excellent speed skating 1500 m event.Provide a model for speed skating coaches that can accurately predict the time spent by athletes in competition,and provide theoretical support for objectively evaluating the athletes’ competitive level,improving the pertinence of training plans,improving the scientific training level of coaches and the arrangement of competition personnel.Research Methods: Literature,Experiment,Mathematical Statistics,BP neural network.The results of the study:（1）Through the correlation analysis and principal component analysis between the indicators obtained by the 30-second Wingate test and the athletes’ competition time,the indicators affecting the competition time of the male and female athletes were as follows: the relative value of maximum power,the minimum power and the final power;the maximum power,maximum speed and total power.（2）According to the linear regression theory,the time-consuming multiple linear regression models of the men’s and women’s athletes in the excellent spe ed skating 1500-meter event in China are: Y=134.654-0.010X2-0.035X6-0.005X9,Y=143.686-0.015X1-0.047X8-1.789E-5X10.（3）The BP neural network prediction model for the men’s and women’s athletes in China’s excellent speed skating 1500 meters event is: 10 neurons in the input layer,1 layer in the implicit layer,10 nodes in the implicit layer,and 1 layer in the output layer.（4）In the form of linear regression combined with BP neural network,a speed skating game time prediction model with 3 input layer neurons,1 output layer,1 hidden layer and 1 node number of the implicit layer is 10 is constructed（5）The three types of prediction models constructed are: multiple linear regression model,BP neural network model and linear regression combined with BP neural network model,and their prediction errors for men’s and women’s athletes are within the acceptable range.Linear regression combined with BP neural network models has the lowest average prediction error value.Research conclusions:（1）The indicators obtained by the 30 s Wingate test can be used to predict the time spent by athletes in the speed skating 1500-meter event in China.（2）The prediction model built by using linear regression combined with BP neural network is more accurate.