基于 ARIMA-ERNN组合模型预测我国甲肝发病率

摘要

Aim:To compare the effect of ARIMA model , BPNN model and ARIMA-ERNN combination model in pre-diction on incidence of hepatitis A in China , and compare the predictive effect among them .Methods: The data of inci-dence of hepatitis A from January 2004 to December 2015 in China were collected and SPSS 17.0 and Eviews 8.0 were used to construct ARIMA model , Matlab 8.0 was used to establish BPNN model and ARIMA-ERNN combination model .At the same time, the data in 2015 was used to evaluate the effect of prediction .Results: The MRE, MER, MSE,RMSE, MAE fitted and forecasted by ARIMA-ERNN combination model were lower than those of ARIMA model and BPNN model , and the MRE was lower than 5%.Conclusion:The ARIMA-ERNN combination model for forecasting the incidence of hep-atitis A is superior to the single ARIMA model and BPNN model .%目的：比较ARIMA模型、BPNN模型和ARIMA-ERNN组合模型在我国甲肝发病率预测中的应用效果，探讨预测甲肝发病率的优化模型。方法：收集2004年1月至2015年12月我国甲肝发病率资料，用SPSS 17．0和Eviews 8．0建立ARIMA模型，用Matlab 8．0建立BPNN模型和ARIMA-ERNN组合模型，并对模型的预测效果进行评价。结果：ARIMA-ERNN组合模型拟合及预测的MRE、MER、MSE、RMSE、MAE均小于ARIMA模型和BPNN模型，其MRE均小于5％。结论：ARIMA-ERNN组合模型对我国甲肝发病率的拟合及预测效果优于ARIMA模型和BPNN模型。