半变系数模型约束PLS估计的渐近正态性(英文)  

Asymptotic Normality of Constrained Profile Least Squares Estimation on Semivarying Coefficient Models

作　　者：吕士钦[1] 张日权[2] 卢准炜[1] 

机构地区：[1]太原理工大学数学系,太原030024 [2]山西大同大学数学系,大同037000

出　　处：《工程数学学报》2009年第2期315-319,共5页Chinese Journal of Engineering Mathematics

基　　金：The Natural Science Foundation of Shanxi Province (2006011006)

摘　　要：半变系数模型已经获得了广泛的研究和应用,近几年,人们提出许多方法来估计其函数系数和常系数。在PLS方法基础上,本文给出半变系数模型模型在线性随机约束条件下的估计,并证明了常系数和函数系数估计的渐近正态性。Semivarying coefficient models have been discussed in many papers and applied widely in many fields. In recent years, many approaches are developed to estimate the unknown parameters and the coefficient functions. Based on the profile least squares method, the constrained profile least squares estimation on semivarying coefficient models under linear stochastic constraints is developed in this paper. Asymptotic normalities of the estimation on parametric component and nonparametric component are investigated.

关 键 词：变系数模型 半变系数模型 渐近正态性 

分 类 号：O212.7[理学—概率论与数理统计]