广义回归神经网络在乙肝发病数时间序列预测中的应用

摘要

To explore the practical value of general regression neural networks ( GRNN) in forecasting the incidence number of hepatitis B (HB) , we make use of HB incidence number information from statutory report of mainland of China from 2005 to 2011 and build respectively the GRNN model and the back propagation neural networks (BPNN) model. Results demonstrate that the mean average error (MAE), mean average percentage error (MAPE) and root mean square error (RMSE) of the values fitted and predicted by the GRNN are all lower than those obtained from BPNN. This result indicates that the GRNN has better applied value in forecasting the incidence of HB.%利用2005年至2011年中国内地法定报告的乙肝发病数资料分别建立广义回归神经网络模型以及传统的BP神经网络模型,探讨广义回归神经网络在乙肝发病预测中的实用价值.结果显示,广义回归神经网络拟合及预测结果的平均绝对误差,平均相对误差以及均方误差均小于BP神经网络.该结果提示,广义回归神经网络在乙肝发病数预测中具有较好的应用价值.