应用SARIMA和ETS模型对湖南省肾综合征出血热发病趋势的预测

Prediction of the Incidence of Renal Syndrome Hemorrhagic Fever in Hunan Province Using SARIMA and ETS Models

目的探讨比较SARIMA模型和ETS模型在湖南省肾综合征出血热的发病预测的应用,为肾综合征出血热的防控提供依据。方法利用2005-2014年肾综合征出血热月度发病数据建立SARIMA模型和ETS模型,并通过模型预测2015年1~12月的肾综合征出血热发病数,用均方根误差(root mean square error, RMSE)和绝对百分比误差(absolutc percentage error, MAPE)作为评价指标。结果 SARIMA(1,0,0)(3,0,0)_(12)模型是肾综合征出血热发病趋势的最优拟合预测模型,SARIMA模型的MAPE为13.18%,低于ETS模型31.14%,SARIMA模型的RMSE为16.27%也低于ETS模型的25.88%。结论 SARIMA(1,0,0)(3,0,0)_(12)模型模拟拟合效果较好,预测结果可为今后肾综合征出血热的预防和控制提供理论支持。

Objective To investigate the application of comparing the SARIMA model and ETS model in predicting the incidence of renal syndrome hemorrhagic fever in Hunan Province,and to provide a basis for the prevention and control of renal syndrome hemorrhagic fever.Methods The SARIMA model and ETS model were established based on the monthly incidence data of hemorrhagic fever with renal syndrome from 2005 to 2014.The models were used to predict hemorrhagic fever incidence with renal syndrome from January to December 2015.Root mean square error(RMSE)and absolute percentage error(MAPE)were used as the predicted evaluation indexes.Results The SARIMA(1,0,0)(3,0,0)12 model is the best-fit prediction model for the trend of hemorrhagic fever with renal syndrome.The MAPE of the SARIMA model is 13.18%lower than the ETS model of 31.14%,the RMSE of the SARIMA model is 16.27%lower than the ETS model of 25.88%.Conclusion The SARIMA(1,0,0)(3,0,0)12 model has a good simulation fit.The prediction results can provide theoretical support for preventing and controlling hemorrhagic fever with renal syndrome in the future.