两种模型对重庆市江北区水痘发病数预测效果的比较  

Comparison between two prediction models on the number of reported varicella cases in Jiangbei District,Chongqing

作　　者：任懿 凌娟 罗淑星 熊丹 牟博媛 熊华利 REN Yi;LING Juan;LUO Shuxing;XIONG Dan;MOU Boyuan;XIONG Huali(Center for Disease Control and Prevention of Jiangbei District,Chongqing 400000,China;不详)

机构地区：[1]重庆市江北区疾病预防控制中心,重庆400000 [2]重庆市荣昌区卫生健康委员会

出　　处：《中国国境卫生检疫杂志》2025年第1期27-31,共5页Chinese Journal of Frontier Health and Quarantine

基　　金：重庆市江北区科卫联合医学科研项目(2023-277-1-26)。

摘　　要：目的使用季节性的自回归综合移动平均(autoregressive integrated moving average,ARIMA)模型和灰色模型[GM(1,1)]拟合重庆市江北区水痘病例的报告数,比较两种模型的预测效果。方法使用2017—2023年重庆市江北区各级医疗机构上报中国疾病预防控制中心传染病信息系统,现住址为“重庆市江北区”的水痘病例,采用SPSS 22.0软件拟合季节性ARIMA模型,用Excel 2013拟合GM(1,1)模型,用平均绝对百分比误差(mean absolute percentage error,MAPE)评价两种模型的拟合效果,以最优模型预测2024—2026的水痘病例报告数。结果2017—2023年重庆市江北区水痘病例报告数分别为911、1412、1542、658、818、674、692例,全年有两个高峰期,第1个高峰期在6月,报告病例833例,占13.17%;第2个高峰期在11月,报告病例数最多,939例,占14.00%。GM(1,1)灰色模型(MAPE=18.56%,后验差比C=0.339,小误差概率p=0.714)预测2024—2026年水痘发病数分别为383、235、194例。结论GM(1,1)模型更适用于预测重庆市江北区水痘病例报告数,预测显示江北区水痘病例的报告数持续下降。Objective To conduct autoregressive integrated moving average model(ARIMA)and the grey model(1,1)[GM(1,1)]by using the number of varicella reported cases in Jiangbei District,Chongqing,compare the predictive performance of two models.Methods The data on reported varicella cases were collected from infectious disease information system of the Chi-nese Center for Disease Control and Prevention from 2017 to 2023,which were reported by various medical institutions with the current address as"Jiangbei District,Chongqing".The seasonal ARIMA model was fitted by SPSS 22.0 software,and the GM(1,1)model was fitted by Excel 2013.The fitting performance of the two models was evaluated by the mean absolute percentage error(MAPE),and the optimal model was used to predict the number of varicella reported cases from 2024 to 2026.Results The reported varicella cases number in Jiang Bei District,Chongqing from 2017 to 2023 were 911,1412,1542,658,818,674 and 692 cases,respectively.There were two peak periods throughout the year,with the first peak in June,reported833 cases(13.17%),and the second peak in November,reported the highest number of cases,939(14.00%).The GM(1,1)gray model(MAPE=18.56%,posterior difference ratio C-value=0.339,probability of small error p-value=0.714)predicted the number of varicella cases to be 383,235,and 194 from 2024 to 2026,respectively.Conclusion The GM(1,1)model was more suitable for predicting the number of reported varicella cases in Jiangbei District,Chongqing,and the prediction indicated a continuous decline in the number of reported varicella cases.

关 键 词：自回归综合移动平均模型 灰色模型 水痘 预测 

分 类 号：R183.3[医药卫生—流行病学]