摘要
目的应用季节指数法和差分自回归移动平均模型(autoregressive integrated moving average model,ARIMA)建立传染病周预测模型。方法利用2008-2011年江苏省感染性腹泻周报告发病数资料,运用SPSS软件计算出季节指数、Eviews软件建立ARIMA模型,最后进行预测精度评价。结果江苏省感染性腹泻周发病数的预测模型为:Yt=St.(-0.335 6DYt-2'-0.206 1DYt-3'+0.572 2DYt-4'+0.174 5DYt-5'+μt+0.478 5μt-1-0.197 5μt-2+0.560 8μt-4+Yt-1'),该模型回代分析的平均误差率为12.44%,对2012年各周发病数进行预测也获得了较好的预测效果。结论运用季节指数法和ARIMA模型可以建立传染病的周预测模型,用于疫情的短期预测和动态分析。
Objective To apply season index method and autoregressive integrated moving average model（ AR1- MA ） to weekly prediction of infectious disease incidence. Methods The data of infectious diarrhea weekly incidence from 2008 to 2011 in Jiangsu Province were used to fit the prediction model by SPSS and Eviews software. Afterwards, the pre- diction model was used to forecast and analyzed the forecast result. Results The model of infectious diarrhea weekly inci- dence in Jiangsu Province was ：Yt=St.（-0.335 6DYt-2＇-0.206 1DYt-3＇＋0.572 2DYt-4＇＋0.174 5DYt-5＇＋μt＋0.478 5μt-1-0.197 5μt-2＋0.560 8μt-4＋Yt-1＇）, which had promising predicting effect. Conclusions The model of weekly pre- diction of infectious disease by joint application of season index method and ARIMA model can be used for short-term fore- casting and dynamic analysis of infectious diseases.
出处
《中华疾病控制杂志》
CAS
北大核心
2013年第8期718-721,共4页
Chinese Journal of Disease Control & Prevention
基金
江苏省"十二.五"科教兴卫工程(ZX201109)
江苏省科教兴卫工程项目(RC2011085)
江苏省疾病预防控制中心科教兴业工程重点人才项目(JKRC2011010)
江苏省预防医学科研项目(YZ201020)
关键词
传染病
预测
腹泻
Communicable disease
Forecasting
Diarrhea
