Prediction Algorithm of Physical Education Achievement Based on BP Neural Network

Achievement prediction is one of the important contents of the research on future development strategy of competitive sports. In this study, take the men's 800 meters final results in 2008 as an example, use BP neural network algorithm to explore the method of achievement prediction on competitive sports. Firstly, expounds the basic principle of BP neural network algorithm, and establish the BP neural network structure diagram; Then, introduces the training process and establish process of BP neural network model, and construct the training process flow chart; Finally, begin the prediction for the achievements of men's 800m based on BP neural network, and through error analysis to test the accuracy of the prediction. BP neural network has the characteristics of massively parallel, distributed processing, self-organizing, self-learning, also has the ability of nonlinear mapping, adaptive ability, generalization ability and fault tolerant ability.