Predictive Model of BP Artificial Neural Network for HFRS Incidence Based on Hyperbolic Tangent Function
基于双曲正切函数HFRS发病率的BP神经网络预测模型

This paper used three layer BP neural network based on Hyperbolic Tangent function as activation function with the number of hidden node six. Presented three nonlinear time series predictive model of BP artificial network for HFRS incidence and prepared them. ANN predictive model III is best, simpler, more effective with high prediction precision. This method can be applied to forecast for disease incidence or mortality.