四川省社会消费品零售额的ARIMA预测模型

摘要

Based on analyzing monthly total sales of consumer goods from 2008 to 2009 in Sichuan Province,China,the author tries to set up ARIMA model.The model＇s forecasting ability has been measured by MAPE（mean absolute percent error）.Secondly,this paper forecasts monthly total sales of consumer goods of 2010 in Sichuan,and compares the predictive value and the actual value.The result indicates that the model provides an excellent forecasting,and the model is fit for forecasting the total sales of consumer goods in Sichuan.Last,some corresponding policy advices have been put forward.%通过分析四川省2008年至2009年每月的社会消费品零售额,建立了ARIMA模型。借助于MAPE（平均绝对百分比误差）度量了模型的预测能力。其次,文章利用所建模型对四川省2010年每月的社会消费品零售额进行预测,并与其实际值进行比较。结果表明,模型的预测精确度非常高,适合于四川省社会消费品零售额的预测。最后,提出了一些相应的政策建议。