SARIMAX模型在肺结核疫情预测中的应用  

Application of SARIMAX model in forecasting epidemic trend of pulmonary tuberculosis

作　　者：张琪 丁焕 陈晓燕[1] 许祝平[1] ZHANG Qi;DING Huan;CHEN Xiao-yan;XU Zhu-ping(Wuxi Municipal Center for Disease Control and Prevention,Jiangsu Wuxi 214023,China)

机构地区：[1]无锡市疾病预防控制中心,江苏无锡214023

出　　处：《江苏预防医学》2020年第6期623-625,共3页Jiangsu Journal of Preventive Medicine

基　　金：江苏省卫生计生委科研课题(Y2018012);无锡市医学重点学科项目(ZDXK009)。

摘　　要：目的探讨SARIMAX模型在肺结核疫情趋势预测中应用的可行性。方法以2013—2018年无锡市肺结核月发病数及气象数据为基础,分别建立SARIMA模型和SARIMAX模型,根据RMSE和MAPE对比两模型的预测效果。结果最终建立SARIMA(1,0,2)(1,0,1)12模型,前瞻性预测误差RMSE和MAPE分别为35.56%、18.28%;SARIMAX(1,0,2)(1,0,1)12模型,两值分别为27.83%、15.08%。结论SARIMAX(1,0,2)(1,0,1)12模型具有更好的拟合优度和预测精度,可用于肺结核疫情趋势预测。Objective To explore the feasibility of SARIMAX model in predicting the epidemic trend of pulmonary tuberculosis.Methods SARIMA and SARIMAX model were fitted by time series of monthly reported pulmonary tuberculosis case numbers from 2013 to 2018 in Wuxi city.The prediction results of the two models were compared according to RMSE and MAPE.Results SARIMA(1,0,2)(1,0,1)12 model was established with prospective RMSE of 35.56%and MAPE of18.28%,respectively.SARIMAX(1,0,2)(1,0,1)12 model was established with prospective RMSE of 27.83%and MAPE of15.08%,respectively.Conclusion SARIMAX(1,0,2)(1,0,1)12 model can be used to forecasting the epidemic trend of pulmonary tuberculosis with better goodness of fit and prediction accuracy.

关 键 词：SARIMAX模型 肺结核 预测 

分 类 号：R521[医药卫生—内科学]