自回归滑动平均混合模型在甘肃省麻疹发病预测中的应用

目的 探讨应用自回归滑动平均混合模型(Autoregressive Integrated Moving Average Model,ARIMA)预测甘肃省麻疹发病率的可行性.方法 应用SPSS 13.0软件对甘肃省1995～2006年麻疹逐月发病率进行ARIMA建模拟合;按照残差不相关原则、简洁原则、赤迟准则与贝叶斯准则建立ARIMA麻疹预测模型,用2007年分月发病数检验模型.结果 模型ARIMA(1,0,0)(0,1,1)1>(不含常数项)所有参数都通过统计学检验,残差序列是白噪声,拟合优度相对最好,参数间也无明显相关性(r=0.069).结论 ARIMA模型很好地模拟和预测了麻疹既往发病周期性以及各月发病数,将其应用于甘肃省麻疹发病预测是可行的.

Study on the Feasibility for ARIMA Model Application to Predict Measles Incidence in Gansu Province

Objective To explore the application of seasonal time series ARIMA model in prediction of measles incidence in Gansu province.Methods SPSS13.0 software was used to construct the ARIMA model based on the monthly measles incidence of Gansu province from Jan.1995 to Dec.2006,with consideration of residual un-correlation and concision.AIC and BIC were used to confirm the fitness of model.The model was tested by the monthly incidence in 2007.Results ARIMA(1,0,0)(0,1,1)12(not with)model was constructed,with signi...