SPORTS INJURY PREDICTION MODEL BASED ON DYNAMIC SAMPLING AND TRANSFER LEARNING

Accurate prediction of vulnerable parts can effectively avoid the injury caused by intense exercise. In order to accurately predict the degree of sports injury, it is necessary to analyze sports injury accurately. Therefore, the design of sports injury prediction model based on dynamic sampling and transfer learning is proposed. The sports injury prediction model based on dynamic sampling and transfer learning describes the influencing factors of vulnerable parts of athletes in high-intensity sports. By calculating the regression function of vulnerable parts and human parts, the functional relationship between the exercise intensity of human body parts and exercise intensity is obtained. Combined with the theory of dynamic sampling and transfer learning, the prediction model of human vulnerable point is established, and the reliability of the prediction model is obtained, and the prediction model of vulnerable point of human body in optimal motion state is given. The simulation results show that the model can accurately predict the situation of sports injury.