应用求和自回归移动平均模型预测福建省甲型病毒性肝炎疫情

Application of ARIMA model in prediction of hepatitis A incidence in Fujian province

目的建立福建省甲型病毒性肝炎(甲肝)分月发病数预测预警的求和自回归移动平均(ARIMA)时间序列模型。方法利用SAS 9.0软件的PROC ARIMA综合软件包对《疾病监测信息报告管理系统》收集的福建省2004-2010年甲肝分月发病数序列进行ARIMA模型的建模与分析。结果福建省2004-2010年甲肝分月发病数序列既含有长期的递减趋势又含有以年为周期的季节效应,经1阶12步差分后为平稳非白噪声序列,拟合的相对最优模型为ARIMA[(12),(1,12),1]。结论拟合甲肝的相对最优ARIMA模型,具有预测和预警的实际应用价值。

Objective To establish the ARIMA model for the prediction of monthly incidence of hepatitis A in Fujian province.Methods The curve of monthly cases of hepatitis A in Fujian province from 2004 to 2010 collected from Disease Information Reporting System was analyzed and ARIMA model was established by using ARIMA procedure of SAS 9.0.Results The case curve was not only with a long-term descending trend but also with annual seasonality.The curve became stable and non-white noise series after 1-lag and 12-lag differences had been taken.The relative optimum fitting model was ARIMA [(12),(1,12),1].Conclusion Fitting a relative optimum model for hepatitis A has practical use in prediction and alert of the disease.