武汉市手足口病发病预测中3种时间序列模型的比较研究  

A comparative study of three time series models in prediction of hand,foot and mouth disease in Wuhan City

作　　者：万源 孔德广[2] 罗俊[1] WAN Yuan;KONG De-guang;LUO Jun(Information Department,Wuhan Center for Disease Control and Prevention,Wuhan Hubei,430024,China;Institute of Infectious Disease Control,Wuhan Center for Disease Control and Prevention,Wuhan Hubei,430024,China)

机构地区：[1]武汉市疾病预防控制中心信息科,湖北武汉430024 [2]武汉市疾病预防控制中心传染病防治所,湖北武汉430024

出　　处：《职业与健康》2021年第22期3098-3101,共4页Occupation and Health

基　　金：武汉市卫生计生委科研课题(WG16B03)。

摘　　要：目的比较3种时间序列分析模型在武汉市手足口病发病预测中效果,探寻预测效果最佳的模型。方法基于2009—2019年《传染病监测系统》中武汉市手足口病月发病率数据,分别构建指数平滑模型、季节趋势模型和自回归滑动平均混合模型,比较3种模型预测值与实际值的平均相对误差(average relate error,MRE)。结果构建的3种模型分别为简单季节性指数平滑模型、季节趋势模型Y_(t)=(9.099×1.003t)×S_(i)、自回归滑动平均混合模型ARIMA(1,0,1)(0,1,1)12;2019年1—6月,3种模型的平均绝对误差(mean absolute error,MAE)分别为16.81%、11.76%、17.36%。结论季节趋势模型在武汉市手足口病的短期发病预测中具有更好的预测效果。Objective To compare the effects of three time series analysis models in predicting the incidence of hand,foot and mouth disease in Wuhan City,explore the model with the best predictive effect. Methods Based on the data on the monthly incidence of hand,foot and mouth disease in Wuhan City from the Infectious Disease Surveillance System,the exponential smoothing model,seasonal trend model and autoregressive integrated moving average model were constructed respectively,and the average relative errors(MRE) between the predicted values of the three models and the actual values were compared. Results The three models constructed were simple seasonal exponential smoothing model,seasonal trend model Y_(t)=(9.099 ×1.003 t) ×S_(i),and autoregressive integrated moving average model ARIMA(1,0,1)(0,1,1)12. From January to June 2019,the mean absolute error(MAE) of the three models was 16.81%,11.76% and 17.36%,respectively. Conclusion The seasonal trend model has a better predictive effect in the short-term incidence prediction of hand,foot and mouth disease in Wuhan City.

关 键 词：手足口病 指数平滑模型 季节趋势模型 自回归滑动平均混合模型 

分 类 号：R181.2[医药卫生—流行病学]