ARIMA乘积季节模型在如东县戊型肝炎发病预测中的应用

ARIMA product season model and its application on the prediction of incidence of hepatitis E

目的通过分析如东县2005-2013年如东县戊型肝炎(戊肝)月发病数据,了解戊肝的发病趋势和季节性特征。探讨ARIMA模型在戊肝发病预测方面的应用,为戊肝的早期预警提供决策依据。方法应用SPSS 20.0软件对2005-2013年如东县戊肝月发病数据进行模型拟合,建立ARIMA模型,用模型对2014年戊肝逐月发病数进行预测分析,并评估模型预测效果。结果 ARIMA乘积季节模型较好地拟合了既往戊肝的实际发病序列,构建ARIMA(1,0,0)×(1,1,0)12模型,残差序列通过了白噪声检验(P>0.05),模型决定系数(R2)为0.92。对2014年各月发病数获得了较好的预测效果。结论 ARIMA模型能较好地模拟如东县戊型肝炎的发病趋势,构建的ARIMA(1,0,0)×(1,1,0)12模型对戊肝发病情况的拟合结果满意,预测效果良好,可用于戊型肝炎疫情的短期预测和动态分析。

Objective To analyze monthly incidence of hepatitis E from January 2005 to December 2013 in Rud- ong county of Jiangsu and to understand the incidence trend and seasonal features of the disease, and to discuss the application of ARIMA model on the prediction of hepatitis E incidence. Methods ARIMA model was es- tablished by model fitting based on monthly incidence data collected from 2005 to 2013 and analyzed by SPSS 20.0 software. Then it was applied to predict the monthly incidence in 2014. Results ARIMA （1, 0, 0） X （1, 1, 0）12 model was successfully established, the residual series were tested by white-noise （P〉0.05） and the R square was 0.92. The model could well predict the monthly incidence in 2014. Conclusions ARIMA product season model is able to predict the incidence of hepatitis E in Rudong of Jiangsu, and can be used for a short-term prediction and a dynamic analysis on hepatitis E.