中文摘要：

目的基于Gramer分解定理，探讨建立确定性趋势模型和ARMA模型的组合模型预测某市艾滋病新发感染率的可行性。方法收集某市40个区县1995～2009年艾滋病月度新发感染资料，建立了指数函数和ARMA（1，0）的组合式模型。结果该组合式模型能较好地拟合当地艾滋病月度新发感染率的变动趋势，且2009年7～12月的预测值符合实际值的变动趋势。结论该组合式模型能够较好地模拟当地艾滋病新发感染率在时间序列上的变动趋势，可以用于短期的预测。

英文摘要：

Objective Based on the Gramer decomposition theorem, this paper explores the combination model of the deterministic model and ARMA model to forecast the HIV/AIDS infection rates change tendency of X city. Methods Collected the data of HIV/AIDS infection of 40 administrative area of X city from 1995 to 2009. Established the index function model and ARMA（ 1,0 ） combined forecasting model. Results The combined model is capable of fitting the AIDS infection rates of new cases of monthly change trend, and in July 2009-December with the actual of the prediction of the trend of change. Conclusion The combined model can exactly fit the incidence of AIDS in time series, which can be used for short-term prediction.