中文摘要：

该文利用SAS统计软件对我国1978-2009年的居民消费水平指数数据进行分析,分别建立了ARIMA模型和Auto-Regressive模型,并给出了反映各个模型拟合精度的AIC值和SBC值,进而确立了一个反映居民消费水平指数变化规律的较优模型.最后,利用该模型对2010年到2014年的全国居民消费水平指数进行了预测.结果表明ARIMA（（2）,2,0）模型在短期预测中达到了较高的精度.

英文摘要：

This paper analyzes the data of China＇s consumption level index from 1978 to 2009 through SAS statistical software.The ARIMA and Auto-Regressive time series model are designed respectively.At the same time,values of AIC and SBC are shown for reflecting the precision of fitting,and then a better model is established which reflects variation regularity of consumption level.Finally,we use it to forecast the consumption level index from 2010 to 2014.The result indicates that ARIMA（（2）,2,0） model has higher accuracy during shor-tem predication.