Predict the College Sports Scores using a Weighted BP-SVR Model

Abstract

A key determinant of college students' physical health is their level of physical accomplishment. As a result, assessment and forecasting of physical accomplishment have grown in significance for both society and university lecturers. Research predicting collegiate athletes' success in sports is currently lacking, and the majority of forecasts were based on imprecise observations. Thus, our goal is to create a model that uses theoretical techniques to forecast the kids' academic success. The Central University of Finance and Economics (CUFE) student basketball performance data was utilized in this study to forecast performance using the back propagation (BP) neural network regression model and the support vector machine regression model (SVR). We thought about integrating the two machine learning prediction techniques to increase accuracy. We created a model using the gradient descent approach, with the real scores acting as dependent factors and the BP and SVR prediction results acting as independent variables, in order to determine the link between the predicted scores of the two models and the real scores. Our methodology, which we call the "BP-SVR weighted prediction model," may be used to predict performance for sports courses other than basketball and is more accurate than only using BP or SVR.