ARIMA乘积季节模型预测永嘉县其他感染性腹泻的流行

Application of multiple seasonal ARIMA model in forecasting the incidence of other infectious diarrhea in Yongjia County

目的评估求和自回归移动平均(ARIMA)乘积季节模型预测其他感染性腹泻流行的可行性。方法利用2005—2014年永嘉县其他感染性腹泻的发病率数据,采用ARIMA模型结合随机季节模型的方法,建立预测其他感染性腹泻流行的ARIMA乘积季节模型,同时用2015年的数据做模型预测效果验证。结果根据模型拟合效果,模型ARIMA(1,1,1)(0,1,1)12的拟合效果为最优,其Ljung-Box检验值为7.796,BIC值为3.602,MAPE值为36.166%,表明模型拟合程度较好;该模型外推验证2015年发病率的预测效果较好,2015年各月发病率的实际值均落在该模型预测值95%可信区间内,且预测值与实际值间依时间变化的趋势也基本一致。结论ARIMA(1,1,1)(0,1,1)12模型能较好地预测永嘉县其他感染性腹泻的流行趋势,对该病的预警具有一定的价值。

Objective To establish the multiple seasonal autoregressive integrated moving average（ ARIMA） model in predicting the incidence of other infectious diarrhea in Yongjia County,and to evaluate the prediction effect of the model.Methods Monthly data of other infectious diarrhea incidence in Yongjia County from January 2005 to December 2014 were obtained from the National Notifiable Disease Surveillance System of China. Combined ARIMA model with stochastic seasonal model was used to establish the multiple seasonal ARIMA model. Monthly incidences of other infectious diarrhea in 2015 were short- term forecasted by the obtained model and compared with the actual data. Results The model ARIMA（ 1,1,1）（ 0,1,1）12was supposed to be the best fitted model. The value of Ljung- Box Q statistic was 7. 796 and Bayesian Information Criteria was 3. 602,and the mean absolute percentage error（ MAPE） value was 36. 166%. The short- term forecasting values in 2015 matched the actual values well,and the actual values were all fall within the 95%confidence interval of the forecasting values. Conclusion The ARIMA（ 1,1,1）（ 0,1,1）12model can be used as a tool for short- term forecasting of the other infectious diarrhea in Yongjia County.