长沙市流行性腮腺炎季节性自回归移动平均模型预测研究

Application of SARIMA Model in the Prediction of Incident Number of Mumps in Changsha

目的采用季节性自回归移动平均(SARIMA)模型预测长沙市流行性腮腺炎发病数。方法收集2005—2015年长沙市报告的流行性腮腺炎病例数据,将2005—2014年数据作为建模数据,将2015年数据作为验证数据,开展SARIMA模型建立与验证研究,并对2016年流行性腮腺炎发病数进行预测。结果 SARIMA(3,0,0)×(1,0,0)_(12)模型可以很好地拟合实际数据,模型的展开式为:Y_t=222.545+1.225Y_(t-1)-0.713Y_(t-2)+0.291Y_(t-3)+0.366Y_(t-12)-0.448Y_(t-13)+0.261Y_(t-14)-0.107Y_(t-15)+a_t。将验证数据与预测数据进行相关性分析,结果显示呈显著性相关(r=0.61,P<0.001)。SARIMA模型预测2016年长沙市全年发病数将达到3 032例,平均月病例数为253例。结论 SARIMA模型可以用于流行性腮腺炎发病数预测,长沙市2016年流行性腮腺炎疫情仍处于高发态势。

Objective To predict the incident number of mumps in Changsha using seasonal autoregressive integrated moving average（ SARIMA） model.Methods We collected the data of incident number of mumps reported in Changsha from2005 to 2015,then,with the data between 2005 and 2014 as the modeling data,and that in 2015 as the validation data,we built and validated a SARIMA model and used it to predict the incident number of mumps in Changsha in 2016.Results SARIMA（ 3,0,0） ×（ 1,0,0）12model could well fit the actual data,the expansion of the model was： Y_t=222.545＋1.225Y_（t-1）-0.713Y_（t-2）＋0.291Y_（t-3）＋0.366Y_（t-12）-0.448Y_（t-13）＋0.261Y_（t-14）-0.107Y_（t-15）＋a_t.The validation data and model prediction data were analyzed,and the results demonstrated that there was significant correlation between them（ r = 0.61,P〈0.001）.SARIMA model predicted that the incident number of mumps occurred in Changsha in the whole year of 2016 would reach 3 032,the average monthly incident number was 253.Conclusion SARIMA model can be used to predict the incident number of mumps.It predicted that the mumps outbreak might still show a trend of high incidence in Changsha in 2016.