重庆市艾滋病发病人数的ARIMA时间序列分析

摘要

Objective To explore the application of ARIMA time series model in predicting the incidence of HIV/AIDS with different routes of infection,verify the feasibility and applicability of the model,and provide more accurate informa-tion for HIV/AIDS prevention and control.Methods The ARIMA model was established based on the HIV/AIDS incidence cases from the first quarter of 2004 to the fourth quarter of 2013 in Chongqing,to predict and evaluate the incidence cases of each quarter of 2014.Results ARIMA(0,1,1)×(0,1,1)4,ARIMA(1,0,0)×(1,1,0)4,ARIMA(0,0,0)×(2,2,0)4were sepa-rately estabilished for the heterosexual behavioral sequence,homosexual behavioral sequence and drug-taking sequence,the pre-dictions were in accordance with the trend of actual incidence,which verified the feasiblity of the model and applicability of sepa-rately establishing model by different routes of infection.Conclusion ARIMA model can be used to simulate the trend of AIDS incidence in time of different routes of infection and make short-term prediction.%目的 探讨按感染途径分类建立ARIMA时间序列模型在艾滋病发病预测方面的应用,验证分析模型的可行性和适用性,为艾滋病防治提供更为精确的信息.方法 利用重庆市2004年第一季度至2013年第四季度的艾滋病发病资料,建立ARIMA模型,对重庆市2014年各季度艾滋病发病人数进行预测和评价.结果 分别为异性性行为序列,同性性行为序列,吸毒序列建立ARIMA(0,1,1)×(0,1,1)4,ARIMA(1,0,0)×(1,1,0)4,ARIMA(0,0,0)×(2,2,0)4模型,预测结果基本符合实际发病人数的变化趋势,验证了模型的可行性及按感染途径分类建模的适用性.结论 ARIMA模型可用于按感染途径分类模拟艾滋病发病在时间上的变化趋势,进行短期的预测.