摘要
目的评价自回归滑动平均模型(ARIMA)、指数平滑模型和季节趋势模型的组合模型对丙型病毒性肝炎(丙肝)发病率预测的效果。方法利用荆州市2007年1月至2015年12月的丙肝逐月发病率作为拟合数据,以2016年1-12月的逐月发病率作为预测数据,分别建立ARIMA、指数平滑模型、季节趋势模型和组合模型,比较4个模型的平均绝对百分比误差(MAPE)、平均误差率(MER)、均方误差(MSE)和平均绝对误差(MAE)。结果 ARIMA模型、指数平滑模型、季节趋势模型和组合模型的拟合的MAPE、MER、MSE、MAE依次分别为18.355%、16.696%、2.417、0.182;15.670%、14.090%、2.047、0.152;15.657%、13.917%、1.974、0.151;15.336%、13.917%、2.016、0.151。4个模型预测的MAPE、MER、MSE、MAE依次分别为14.034%、15.001%、1.185、0.283;13.316%、14.184%、1.112、0.267;10.491%、10.865%、0.834、0.205;12.031%、12.776%、0.992、0.241。结论组合模型的拟合效果优于单一模型,预测效果优于或相当于单一模型。
Objective To evaluate the combination of an autoregressive integrated moving average(ARIMA) model, an exponential smoothing model, and a seasonal trend model for the prediction of hepatitis C incidence. Methods We selected monthly incidences of hepatitis C reported from January 2007 to December 2015 in Jingzhou city for model-fitting data and predicted monthly incidences from January to December 2016 by establishing an ARIMA model, an exponential smoothing model, a seasonal trend model, and a combination model. We compared the mean absolute percentage error(MAPE), mean error rate(MER), mean square error(MSE), and mean absolute error(MAE) of the four models. Results In the four fitted models, MAPE, MSE, MER, and MAE were respectively 18.355%, 16.696%, 2.417, and 0.182; 15.670%, 14.090%, 2.047, and 0.152; 15.657%, 13.917%, 1.974, and 0.151; and 15.336%, 13.917%, 2.016, and 0.151. MAPE, MSE, MER, and MAE predicted by the four models were respectively 14.034%, 15.001%, 1.185, and 0.283; 13.316%, 14.184%, 1.112, and 0.267; 10.491%, 10.865%, 0.834, and 0.205; and 12.031%, 12.776%, 0.992, and 0.241. Conclusions The combination model performed better than the single models in fitting effect, and better than or similar to the single models in prediction effect.
作者
刘天
姚梦雷
黄继贵
毛安禄
陈红缨
黄淑琼
杨雯雯
蔡晶
吴然
Liu Tian;Yao Menglei;Huang Jigui;Mao Anlu;Chen Hongying;Huang Shuqiong;Yang Wenwen;Cai Jing;Wu Ran(Jingzhou Municipal Center for Disease Control and Prevention,Jingzhou 434000,Hubei, China;Hubei Provincial Center for Disease Control and Prevention,Wuhan 430000,Hubei,China)
出处
《中国疫苗和免疫》
北大核心
2018年第6期675-680,共6页
Chinese Journal of Vaccines and Immunization
基金
湖北省卫生计生委创新团队项目(WJ2016JT-002)
关键词
ARIMA
指数平滑模型
季节趋势模型
组合模型
预测
ARIMA
Exponential smoothing model
Seasonal trend model
Combination model
Prediction
作者简介
通信作者:姚梦雷,Email:jzcrbs@163.com
