深圳市登革热新发病例及蚊媒密度SARIMA预测模型分析  

Analysis of a predictive SARIMA model for new dengue fever cases and mosquito vector density in Shenzhen

作　　者：林良强[1] 孔东锋[1] 项兰斌 陈志高[1] 秦彦珉[1] 刘阳[1] 李剑锋[1] 李苑[2] LIN Liangqiang;KONG dongfeng;XIANG Lanbin;CHEN Zhigao;QING Yanmin;LIU Yang;LI Jianfeng;LI Yuan(Shenzhen Center for Disease Control and Prevention,Shenzhen,Guangdong 518000,China;Bao'an District Center for Disease Control and Prevention,Shenzhen,Guangdong 518101,China)

机构地区：[1]深圳市疾病预防控制中心,广东深圳518000 [2]深圳市宝安区疾病预防控制中心,广东深圳518101

出　　处：《中华卫生杀虫药械》2025年第2期210-216,225,共8页Chinese Journal of Hygienic Insecticides and Equipments

基　　金：2024年深圳市医学研究专项资金项目(B2404002)。

摘　　要：目的运用季节性差分求和自回归移动平均模型(SARIMA)分析和预测深圳市登革热发病和白纹伊蚊诱蚊诱卵指数(MOI)的变化趋势。方法收集2011年1月1日至2018年12月31日每月登革热发病和MOI数据,运用SPSS 20.0和Eviews 9.0统计软件拟合SARIMA模型,并利用2018年1—12月的实际观测值与预测值进行分析比较,评估模型预测效果。结果深圳市登革热发病率与MOI均呈现明显的周期性和季节性,利用SARIMA(0,0,1)(0,1,1)12和SARIMA(1,1,0)(0,1,1)12分别预测了每月登革热发病率及MOI水平。2种不同SARIMA模型产生的预测值与实际值基本一致,其中登革热发病率预测模型中平均绝对误差百分比(MAPE)为175.74%,平均绝对误差(MAE)为10.59;而MOI预测模型中MAPE为41.63%,MAE为1.538。结论SARIMA(0,0,1)(0,1,1)12和SARIMA(1,1,0)(0,1,1)12模型可用于预测深圳市登革热的发生和传播风险。Objective To analyze and predict the incidence trend of dengue fever and mosquito ovitrap index(MOI)for Aedes albopictus in Shenzhen using the seasonal auto-regressive integrated moving average(SARIMA)model.Methods Monthly data of dengue fever and MOI were obtained from 1 January 2011 to 31 December 2018.The SARIMA model was performed with SPSS 20.0 and Eviews 9.0 statistical software.The data collected from January to December 2018 were used to validate the model through the comparison between the actual and the predicted value.Results The incidence of dengue fever and MOI showed obvious periodicity and seasonality in Shenzhen.The SARIMA(0,0,1)(0,1,1)12 and the SARIMA(1,1,0)(0,1,1)12 were developed to predict the monthly incidence of dengue fever and MOI,respectively.The predictive values generated by the two different SARIMA models were generally consistent with the actual values;with the mean absolute percentage error(MAPE)of 175.74% and the mean absolute error(MAE)of 10.59 in the incidence model,and the MAPE of 41.63% and the MAE of 1.538 in the MOI model.Conclusion Both SARIMA(0,0,1)(0,1,1)12 and SARIMA(1,1,0)(0,1,1)12 models can be applied to predict the occurrence and transmission of dengue fever in Shenzhen.

关 键 词：登革热 诱蚊诱卵指数(MOI) SARIMA模型 时间序列分析 

分 类 号：R373.33[医药卫生—病原生物学]