SARIMA模型在苏州市肺结核发病预测中的应用  

Application of SARIMA model in the prediction of tuberculosis incidence in Suzhou

作　　者：傅颖 张晓龙[1] 蒋骏[1] 李云[1] 王斐娴 FU Ying;ZHANG Xiaolong;JIANG Jun;LI Yun;WANG Feixian(Department of Tuberculosis Control and Prevention,Suzhou Center for Disease Control and Prevention,Suzhou,Jiangsu 215004,China)

机构地区：[1]苏州市疾病预防控制中心结核病防制所,江苏苏州215004

出　　处：《中国热带医学》2020年第4期339-342,共4页China Tropical Medicine

基　　金：苏州市科技计划项目(No.SYS201661);苏州市“科教兴卫”青年科技项目(No.kjxw2014034)。

摘　　要：目的建立苏州市肺结核发病的SARIMA模型并预测发病,为苏州市肺结核防控提供参考。方法收集结核病信息管理系统(新)中苏州市2010年1月—2018年12月肺结核月发病数,通过时间序列分析建立SARIMA模型并预测苏州市2019年肺结核的发病情况。结果苏州市肺结核发病数具有明显的季节周期性,每年的发病最高峰为5月,发病最低谷为2月。苏州市肺结核发病数的最佳拟合模型为SARIMA(0,1,1)×(0,1,1)12,AIC=9.590,SBC=9.644,模型参数均具有统计学意义,模型残差为白噪声序列,模型的预测值与实际值平均绝对百分比误差MAPE=7.943%,模型预测精度较高。预测苏州市2019年肺结核发病数为3467例,月发病数平均值为289例,发病水平较2018年略有下降。结论SARIMA(0,1,1)×(0,1,1)12模型能较好拟合出苏州市肺结核发病数的时间变化趋势,可应用于苏州市肺结核月发病数的短期预测。Objective To establish a SARIMA model and predict the incidence of tuberculosis in Suzhou,and we provide reference for prevention and control of tuberculosis in Suzhou.Methods The monthly incidence of tuberculosis in Suzhou was collected from January 2010 to December 2018 in tuberculosis information management system(new).Through time-series analysis,a SARIMA model was established to predict the incidence of tuberculosis in Suzhou in 2019.Results The incidence of tuberculosis in Suzhou is of distinct seasonality.The peak month is May,and the trough is February.Best fitting model for the incidence of tuberculosis in Suzhou is SARIMA(0,1,1)×(0,1,1)12,AIC=9.590,SBC=9.644.Model parameters are statistically significant,and model residual is white noise.Mean absolute percentage error(MAPE)between predictive and actual values is 7.943%.The model is of high prediction accuracy.It is predicted that the number of tuberculosis cases in Suzhou will be 3467 in 2019,and the average number of monthly incidence is 289.The incidence level of 2019 is slightly lower than 2018.Conclusions SARIMA(0,1,1)×(0,1,1)12 model exactly fitted the changes in the number of tuberculosis cases in Suzhou over time,and can be used for short-term prediction of the monthly incidence of tuberculosis in Suzhou.

关 键 词：肺结核 SARIMA模型 预测 发病 

分 类 号：R521[医药卫生—内科学]