基于ARIMA时间序列的广西百色市艾滋病流行趋势分析  

Analysis of AIDS epidemic trend in Baise city of Guangxi based on ARIMA time series

作　　者：李阳 陈坚[3] 廖建英[2] 谢艺红 Li Yang;Chen Jian;Liao Jianying;Xie Yihong(Guangxi Medical University,Nanning 530000,Guangxi,China;Youjiang Medical University for Nationalities,Baise 533000,Guangxi,China;Baise Centre for Disease Control and Prevention,Baise 533000,Guangxi,China)

机构地区：[1]广西医科大学,广西南宁530000 [2]右江民族医学院,广西百色533000 [3]广西百色市疾病预防控制中心,广西百色533000

出　　处：《右江民族医学院学报》2023年第1期134-138,共5页Journal of Youjiang Medical University for Nationalities

基　　金：百色市科学研究与技术开发计划项目(百科20184201,百科20222037)。

摘　　要：目的分析广西百色市2011年至2021年艾滋病报告病例的分布特征,建立ARIMA时间序列模型,对艾滋病报告病例数进行短期预测,为百色市艾滋病防治策略和措施的制定提供科学依据。方法对2011年至2021年中国疾病预防控制信息系统艾滋病监测子系统现住址为百色的报告病例进行描述性分析,并采用Eviews 10.0统计学软件建立ARIMA时间序列模型进行预测。结果百色市2011年至2021年艾滋病报告病例数总体呈上升趋势,病例有明显的地区聚集性,东南部高发,重点分布在右江区、田阳区、平果市、田东县、靖西市等地;报告病例无明显季节性。经过数据序列平稳化处理和参数估计,确定ARIMA(2,1,2)为最优拟合模型。模型的平均拟合准确度为85.29%,平均相对误差为14.71%,说明该模型的拟合效果良好。结论百色市艾滋病报告病例数呈逐年上升趋势,且有明显的地区聚集性,ARIMA时间序列模型能较好地预测百色市艾滋病报告病例数,应加强高发地区的防控和危险因素监测,并制定针对性强的综合防控策略和措施。Objective To analyze the distribution characteristics of reported AIDS cases from 2011 to 2021 in Baise city,Guangxi Zhuang Autonomous Region,and establish the model of ARIMA time series to predict the number of reported AIDS cases in the short term,so as to provide scientific basis for the formulation of strategies and measures for AIDS prevention and treatment in Baise city.Methods Descriptive analysis was conducted on the reported cases whose current address was in Baise according to the AIDS surveillance subsystem of China Information System for Disease Control and Preventionfrom 2011 to 2021.The model of ARIMA time series was established with the statistical software Eviews 10.0 for prediction.Results From 2011 to 2021,the reported AIDS cases in Baise showed an overall increasing trend,with obvious regional clustering and high incidence in the southeast,mainly distributed in Youjiang district,Tianyang district,Pingguo city,Tiandong county,Jingxi city and other places.There was no significant difference in seasonality in the reported cases.After stabilization of data series and estimation of parameter,ARIMA(2,1,2)was identified as the optimal fitting model.The average fitting accuracy of the model was 85.29%,and the average relative error was 14.71%,indicating that the model had a good fitting effect.Conclusion The number of reported AIDS cases in Baise city turns out to be increasing year by year and there is obvious regional aggregation.The model of ARIMA time series can predict the number of reported AIDS cases in Baise city.It is necessary to strengthen the prevention and control and monitoring of risk factors in high-incidence areas,and to formulate well-targeted comprehensive strategies and measures for prevention and control of AIDS.

关 键 词：获得性免疫缺陷综合征 ARIMA模型 时间序列 预测 

分 类 号：R512.91[医药卫生—内科学]