【摘要】 目的探讨利用季节性差分自回归移动平均模型（SARIMA）模型预测南昌市HIV/AIDS发病情况,为该市艾滋病防控提供决策依据。方法基于艾滋病综合防治信息系统中2007年1月至2016年12月HIV/AIDS病例人数建立SARIMA模型,用2017年1—6月HIV/AIDS病例数验证模型的预测效果,然后用建立的模型预测2017年7—12月HIV/AIDS病例数。结果南昌市2007—2016年共报告HIV/AIDS病例数2218例,且HIV/AIDS病例数呈逐年上升趋势,发病具有明显的季节性特征。最终建立的最优模型为SARIMA（0,1,1）×（0,0,1）₁₂,赤池信息准则（AIC）为771.0,平均绝对百分误差（MAPE）为29.6%,且通过Ljung-Box检验,模型残差为白噪声（P=0.686）。模型对2017年1—6月发病数进行预测,均在95%置信区间内,且与实际报告病例数变动的趋势一致,验证了模型合理性。模型预测2017年7—12月HIV/AIDS病例数也符合之前流行趋势。结论 SARIMA模型能较好地拟合南昌市HIV/AIDS按月发病人数的动态变化,可用于HIV/AIDS疫情的短期预测。更多还原

【Abstract】 Objective To explore the use of the seasonal autoregressive integrated moving average（SARIMA） model in predicting the incidence of HIV/AIDS in Nanchang City,and to provide decision-making basis for AIDS prevention and control.Methods The SARIMA model was established based on the number of HIV/AIDS cases reported from January 2007 to December 2016 in the Integrated AIDS Prevention and Control Information System.The prediction effect of the model was verified by the number of HIV/AIDS cases from January to June 2017,and then the model was used to predict the number of HIV/AIDS cases from July to December 2017.Results A total of 2218 HIV/AIDS cases were reported in Nanchang from 2007 to 2016.The incidence of HIV/AIDS showed an increasing trend year by year and had obvious seasonal characteristics.The optimal model finally established was the SARIMA（0,1,1）×（0,0,1）₁₂ with a Akaike Information Criterion（AIC） of 771.0 and a mean absolute percentage error（MAPE） of 29.6%,which passed the Ljung-Box test.The model residual was white noise（P=0.686）.The number of cases predicted by the model between January and June 2017 was within the 95% confidence interval and was consistent with the trend of the number of reported cases,confirming the rationality of the model.In addition,the number of HIV/AIDS cases predicted by the model between July and December 2017 was also consistent with the previous epidemic trend.ConclusionThe SARIMA model fits the dynamic changes of the monthly incidence of HIV/AIDS in Nanchang,and can be used for short-term prediction of HIV/AIDS epidemic.更多还原