Proper

Abstract

Diabetes is seriously harmful to human health. Early detection, early diagnosis and early treatment can reduce the possibility of its complications and mortality. Mathematical prediction model can effectively solve the above problems and provide helpful information for clinical in order to diagnose the disease in a more rapid, effective and accurate way. Therefore, a diabetes predicting model based on BP neural network is proposed, which implements BP neural network training combined with the cross validation to realize the optimal model structure through setting input, BP network structure, the number of nodes in different layers, BP network parameters and other steps. For the Pima Indians Diabetes data set, the optimal BP neural network construction is set to 8-12-2, and the classification accuracy rate of the final test set in the constructed BP neural network prediction model reaches 76.43%. The experimental results show that the classification accuracy of the model is improved and the model performance is good.