自回归求和移动平均乘积季节模型在山东省猩红热发病预测中的应用及效果评价  

Effects evaluation and application on multiple seasonal ARIMA model in predicting incidence of scarlet fever,Shandong province

作　　者：刘维量[1] 马洁 窦伟洁[2] LIU Wei-liang;MA Jie;DOU Wei-jie(Shandong Center for Disease Control and Prevention,Jinan,Shandong,250014,China;不详)

机构地区：[1]山东省疾病预防控制中心,山东济南250014 [2]山东省医药卫生科技信息研究所山东第一医科大学(山东省医学科学院),山东济南250062

出　　处：《预防医学论坛》2020年第10期740-742,745,共4页Preventive Medicine Tribune

基　　金：山东省医药卫生科技发展计划项目(项目编号:2018WS310)。

摘　　要：目的应用自回归求和移动平均(autoregressive integrated moving average model,ARIMA)乘积季节模型分析山东省猩红热季节性时间序列,探讨建立山东省猩红热发病数ARIMA预测模型的可行性。方法应用SAS软件对2012年1月至2018年12月山东省猩红热月报告发病数建立ARIMA乘积季节预测模型,利用2019年1~9月数据进行模型检验,评价预测效果。结果山东省猩红热分月报告发病数存在明显的季节性规律,发病高峰分别为4~6和11月到次年1月。建立的ARIMA(1,0,0)(1,1,0)6乘积季节模型所有参数均通过统计学检验(P<0.05),模型表达式为■■6xt=1/(1-0.081292B)(1+0.82844B6)εt;2019年1~9月预测值发病趋势与实际值基本一致,实际值均在预测值的95%置信区间(95%CI)内。结论ARIMA(1,0,0)(1,1,0)6乘积季节模型预测效果较好,可用于山东省猩红热月报告发病人数的预测。Objective To analyze the seasonal time series by using the multiple seasonal autoregressive integrated moving average model(ARIMA),and explore the feasibility of multiple seasonal ARIMA in predicting the number of scarlet fever in Shandong province.Methods In Shandong province,SAS software was used to establish multiple seasonal ARIMA model based on the monthly incidence data of scarlet fever during January 2012-December 2018.Then the ARIMA model was evaluated by the data from January to September 2019.Results There was a strong seasonal nature in the reported cases of scarlet fever in Shandong,and the annual incidence peak of scarlet fever occurred from April to June and from November to next January.All parameters of ARIMA(1,0,0)(1,1,0)6were statistically tested(P<0.05),and the model equation was■■6xt=1/(1-0.081292B)(1+0.82844B6)εt.The prediction result was quite similar to the actual values of scarlet fever from January to September in 2019,and the actual values of the prediction results were within the 95%confidence interval(95%CI)of the predicted value.Conclusion The multiple seasonal ARIMA(1,0,0)(1,1,0)6 model is better,which can be used to predict the monthly number of scarlet fever cases in Shandong province.

关 键 词：自回归求和移动平均(ARIMA)乘积季节模型 时间序列 预测 猩红热 

分 类 号：R515.1[医药卫生—内科学]