SARIMA模型在猩红热疫情趋势预测中的应用  

Application of SARIMA model in forecasting epidemic trend of scarlet fever

作　　者：张琪[1] 刘文东[2] 吴莹[2] 时影影[2] 郑金鑫[3] 朱叶飞 周明浩[1,2] ZHANG Qi;LIU Wen-dong;WU Ying;SHI Ying-yin;ZHENG Jin-xin;ZHU Ye-fei;ZHOU Ming-hao(Nanjing Medical University, School of Public Health, Nanjing 211166, China)

机构地区：[1]南京医科大学公共卫生学院,南京211166 [2]江苏省疾病预防控制中心 [3]江苏省血吸虫病防治研究所 [4]南京医科大学第二附属医院

出　　处：《江苏预防医学》2017年第6期655-658,共4页Jiangsu Journal of Preventive Medicine

基　　金：重大新发传染病综合防控科技示范工程(BE2015714);江苏省自然科学基金(BK20151595);江苏省临床医学科技专项(BL2014081)

摘　　要：目的探讨SARIMA模型在猩红热疫情趋势预测中的应用,为江苏省猩红热疫情的及时防控提供科学依据。方法以2005—2015年江苏省猩红热月发病数为基础,建立SARIMA模型,对江苏省2016—2018年猩红热疫情趋势进行预测。结果最终建立SARIMA(1,1,0)(0,1,1)12,回代误差为15.47%,前瞻性预测误差为12.98%,预测结果表明,江苏省2016—2018年猩红热疫情依然呈现上升趋势。结论 SARIMA(1,1,1)(0,1,1)12模型具有较好的拟合优度和前瞻性预测精度,可用于猩红热疫情趋势的研究;江苏省猩红热疫情依然严峻,各地应加强疫情防控工作。Objective To explore the feasibility of SARIMA model in forecasting the epidemic trend of scarlet fever;to provide scientific evidence for timely prevention and control of scarlet fever.Methods SARIMA model was fitted by the time series of monthly reported case numbers from2005to2015of scarlet fever in Jiangsu province,the epidemic trend of scarlet fever from2016to2018was forecasted.Results SARIMA(1,1,0)(0,1,1)12model was fitted with the15.47%MAPE and12.98%prospective MAPE.The result forecasted by SARIMA(1,1,0)(0,1,1)12indicated an ascending epidemic trend of scarlet fever in Jiangsu province from2016to2018.Conclusion SARIMA(1,1,0)(0,1,1)12model can be used in forecasting the epidemic trend of scarlet fever with the good prediction accuracy and goodness of fit;the situation of scarlet fever in Jiangsu province is still serious,and the prevention of scarlet fever should be strengthened.

关 键 词：时间序列分析 SARIMA模型 猩红热 预测 

分 类 号：R515.1[医药卫生—内科学]