基于强混合数据的半参数平滑转换回归模型(英文)  

The Strong Mixing Data-Based Semiparametric Smooth Transition Regression Model

作　　者：王成勇[1] 艾春荣[2] 

机构地区：[1]湖北文理学院数学与计算机科学学院,襄阳441053 [2]上海财经大学统计与管理学院,上海200433

出　　处：《应用概率统计》2013年第4期348-362,共15页Chinese Journal of Applied Probability and Statistics

基　　金：supported by Humanity and Social Science Youth Foundation Grants of National Education Bureau(10YJC790247);The National Natural Science Foundation of China(70971082)

摘　　要：本文基于严平稳强混合数据和带确定性趋势的强混合数据序列,推广了文献[20]中提出的半参数平滑转换回归模型.对含于平滑转换函数中的未知光滑有界函数应用级数估计方法,并基于非线性最小二乘估计和级数估计理论证明了模型参数估计量的相合性和渐近正态性等大样本性质,简要讨论了其协方差矩阵的估计以及假设检验问题.最后,应用该模型重新研究了我国年度通货膨胀率的平滑转换结构.In this paper, we generalize the semiparametric smooth transition regression model proposed by Wang （2012a）, to adapt for the strictly stationary strong mixing data and strong mixing data with deterministic trends. The unknown bounded smooth function embedded in the smooth transi- tion function is estimated by series estimator, the consistency and asymptotic normality properties of estimators are proved employing nonlinear least square regression theory and series estimator approach. Variance matrix estimation and hypothesis testing problems are also discussed based on estimated standard errors. The new model is then used to study the annually inflation rates of China.

关 键 词：平滑转换回归模型 级数估计 相合性 渐近正态性性 

分 类 号：O212.7[理学—概率论与数理统计]