自回归移动平均模型在深圳市宝安区结核病疫情预测中的应用  

Application of autoregressive integrated moving average model in the forecasting of tuberculosis epidemic situation in Bao’an District,Shenzhen

作　　者：张娟娟[1] 王云霞[1] 胡方祥 黄佩佩 刘振扬[1] 梅金周 袁青[1] Zhang Juanjuan;Wang Yunxia;Hu Fangxiang;Huang Peipei;Liu Zhenyang;Mei Jinzhou;Yuan Qing(Department of Tuberculosis Control,Bao'an District Chronic Disease Control Hospital,Guangdong Shenzhen 518100,China)

机构地区：[1]深圳市宝安区慢性病防治院结核病防治科,广东深圳518100

出　　处：《新发传染病电子杂志》2022年第1期47-51,共5页Electronic Journal of Emerging Infectious Diseases

基　　金：“十三五”国家科技重大专项(2018ZX10715004);宝安区科技创新局医疗卫生基础研究项目(2018JD047)。

摘　　要：目的 建立自回归移动平均(ARIMA)模型,预测深圳市宝安区结核病月发病情况,为肺结核防控措施的制定提供科学参考。方法 通过中国疾病预防控制系统结核病信息管理系统导出2006-2020年深圳市宝安区月报告患者例数。采用IBM SPSS 20.0统计学软件,以2006年1月至2019年12月的月报告患者例数为基础建立时间序列,构建ARIMA模型,通过对模型的识别、定阶、诊断,筛查出最优模型,利用该模型预测2020年1-12月结核病的月发病情况,通过比较预测值与实际值来评价拟合模型的预测效果。结果 ARIMA(1,0,0)(0,1,1)12模型,其参数均通过统计学检验(P<0.05),残差序列为白噪声序列(P>0.05),R^(2)=0.561,RMSE=22.632,NBIC=6.336,拟合优度相对较好。2020年1-12月的预测值与实际值基本吻合,实际值均落在95%CI内,预测效果较好。结论 ARIMA(1,0,0)(0,1,1)12模型可用于短期预测深圳市宝安区结核病疫情,预测效果较好。Objective To establish an autoregressive integrated moving average (ARIMA) model to predict the monthly incidence of tuberculosis in Bao’an District,Shenzhen,and to provide a scientific reference for the formulation of tuberculosis prevention and control measures.Method The number of patients reported monthly in Bao’an District,Shenzhen from 2006 to 2020 was derived from the tuberculosis information management system of the China Disease Control and Prevention System.Use IBM SPSS 20.0 statistical software to establish a time series based on the number of patients reported monthly from January 2006 to December 2019,construct an ARIMA model,and screen out the optimal model through identification,ranking,and diagnosis of the model.The model is used to predict the monthly incidence of tuberculosis from 1 to 12 in 2020,and the prediction effect of the fitted model is evaluated by comparing the predicted value with the actual value.Result The parameters of the ARIMA(1,0,0)(0,1,1)12 model passed the statistical test (P<0.05),and the residual sequence was a white noise sequence (P>0.05),R^(2)=0.561,RMSE=22.632,NBIC=6.336,the goodness of fit is relatively good.The predicted value from January to December 2020 is basically consistent with the actual value,and the actual value falls within the 95% CI,and the prediction effect is good.Conclusion The ARIMA(1,0,0)(0,1,1)12 model can be used for short-term prediction of tuberculosis epidemic in Baoan District,Shenzhen,and the prediction effect is good.

关 键 词：肺结核 自回归移动平均模型 疫情 预测 

分 类 号：R521[医药卫生—内科学]