加入季节因子的ARIMA模型在白纹伊蚊布雷图指数预测中的应用  

Application of ARIMA model with seasonal factors in prediction of BI of Aedes albopictus

作　　者：韩晓莉[1] 高文[1] 翟士勇[2] 赵勇[1] 宋纪文[1] 王喜明[1] HAN Xiao-li;GAO Wen;ZHAI Shi-yong;ZHAO Yong;SONG Ji-wen;WANG Xi-ming(Center for Disease Control and Prevention of Hebei Province,Shijiazhuang 050021,China;Center for Disease Control and Prevention of Shijiazhuang City,Shijiazhuang 0500211 China)

机构地区：[1]河北省疾病预防控制中心,河北石家庄050021 [2]石家庄市疾病预防控制中心,河北石家庄050021

出　　处：《中华卫生杀虫药械》2021年第6期501-506,共6页Chinese Journal of Hygienic Insecticides and Equipments

基　　金：2021年度河北省医学研究课题计划(编号:20210681)。

摘　　要：目的探讨基于季节因子的差分自回归移动平均(ARIMA)模型在白纹伊蚊布雷图指数(BI)预测中的可行性。方法应用SPSS 19.0将2015—2018年河北省登革热媒介伊蚊监测点布雷图指数法监测数据构建基于季节周期因子的ARIMA模型,进行整体回代评价拟合效果,比较2019年6—9月的预测值与真实值,评价外推效果。结果构建基于季节因子的ARIMA模型方程为:y1=SAF+1/(1+0.551b+0.404B^(2))(1-B)ε_(i)(平稳的R^(2)=0.306,BIC=2.292,平均绝对误差MAE=2.280);2019年6—9月预测值与实际值基本一致,平均绝对百分误差(MAPE)为22.19%。结论基于季节因子的ARIMA模型能够较好地预测河北省白纹伊蚊BI消长趋势,可为蚊媒传染病防控提供参考。Objective To explore the feasibility of seasonal factor-based ARIMA model in the prediction of BI of Aedes albopictus.Methods The ARIMA model based on seasonal cycle factor was constructed by using the data of dengue vector Aedes mosquito monitoring from 2015 to 2018 by SPSS 19.0.The overall back-generation evaluation of the fitting effect was performed,the comparison between the predicted value from June to September of 2019 and the true value was made,and the effect of extrapolation was evaluated.Results The equation for constructing the ARIMA model based on seasonal factors was:y1=SAF+1/(1+0.551b+0.404B^(2))(1-B)ε_(i),stationary R^(2)=0.306,average absolute error MAE=2.280);the predicted value from June to September in 2019 was basically consistent with the actual value,and the mean absolute percentage error(MAPE)was 22.19%.Conclusion The ARIMA model based on seasonal factors can better predict the BI trend of Aedes albopictus in Hebei Province,which provide the reference for the prevention and control of mosquito-borne infectious diseases.

关 键 词：ARIMA模型 季节因子 白纹伊蚊 布雷图指数 预测 

分 类 号：R384.1[医药卫生—医学寄生虫学]