基于NAR模型的商品零售价格指数预测研究

The Prediction of Commodity Retail Price Index Based on the NAR Model

基于样条估计理论,建立了我国商品零售价格指数的NAR预测模型.并应用此模型对我国2004—2007年的价格指数进行了预测分析,结果表明:相对于传统的ARMA模型而言,NAR模型能够更好地捕获我国社会商品零售价格指数的非线性特征,预测精度较高.

Based on spline estimation theory,the nonparametric autoregression forecast model is established on our country＇s Commodity Retail Price Index.And this model is applied to predict the observations 20042007.The computed results show that the nonparametric autoregression model can catch the nonlinear characteristic of our country＇s Commodity Retail Price Index better than the ARMA model.