采用ARIMA模型对河北省保定市手足口病发病的预测分析  

Predictive analysis of ARIMA model on hand,foot and mouth disease(HFMD) of Baoding City of Hebei Province

作　　者：王磊[1] 王春艳[1] 祖文刚[1] 崔立周[1] WANG Lei;WANG Chunyan;ZU Wengang;CUI Lizhou(Baoding City Center for Disease Control and Prevention,Hebei 071000,China)

机构地区：[1]保定市疾病预防控制中心,河北071000

出　　处：《医学动物防制》2018年第9期836-839,共4页Journal of Medical Pest Control

基　　金：河北省卫生厅医学科学研究重点课题(20181504)

摘　　要：目的采用自回归滑动平均模型(autoregressive integrated moving average model,ARIMA)对保定市手足口病发病预测进行模型构建,探索该市传染病预测的方法。方法利用软件SPSS 13.0,对保定市2009-2015年手足口病逐月发病数进行ARIMA建模和拟合,筛出最优模型,对2016年1~10月的发病数进行预测并评价效果。结果 ARIMA(0,1,1)×(1,1,0)12为最适模型,对保定市2009-2015年手足口病发病数进行了很好的拟合,残差是白噪声序列,对2016年1~10月发病数的预测值均落入相应的95%可信区间,再现了手足口病的季节模式。结论 ARIMA模型很好的拟合了保定市手足口病在时间序列上的变化趋势,可用于疫情的动态分析和短期预测。Objective To construct predictive model, with ARIMA（autoregressive integrated moving average model）, of HFMD of Baoding City,to explore the way of the prediction of infectious diseases. Methods The database of HFMD month-situation in Baoding City from 2009 to 2016 was constructed to be fitting using ARIMA with SPSS 13. 0 software,screen out the optimal model,in January 2016 to October to predict the incidence and evaluation of the effect. Results ARIMA（0,1,1） ×（1,1,0）12 for the optimal model of the Baoding City in 2009-2015,HFMD incidence had good fitting,residual is white noise sequence,in January 2016 to October incidence forecasts are falling into the corresponding 95% confidence interval,reproduced the seasonal pattern of HFMD. Conclusion ARIMA model good fitting for HFMD of Baoding City change trend in time series,can be used in the dynamic analysis of epidemic situation and short-term forecasting.

关 键 词：ARIMA模型 手足口病 预测 

分 类 号：R512.5[医药卫生—内科学]