半变系数模型约束PLS估计的渐近正态性(英文)  

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

作　　者：吕士钦[1] 卢准炜[1] 张日权[1] 

机构地区：[1]太原理工大学理学院,山西太原030024

出　　处：《应用数学》2008年第1期149-155,共7页Mathematica Applicata

基　　金：the Nature Science Foundation of Shanxi(2006011006)

摘　　要：变系数模型已获得了广泛的应用,半变系数模型是变系数模型的有效推广,本文给出半变系数模型在线性约束条件下的PLS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 are efficient generalization of the varying-coefficient regression models which have extremely wide applied. In recent years ,many approaches are developed to estimate the unknown parameters and the coefficient functions. Based on the ideals of Fan and Huang （2005）, the constrained profile least squares estimation on semivarying coefficient models under linear constraints is developed in this paper. Asymptotic normalities of the estimation on parametric component and nonpararnetric component are investigated.

关 键 词：变系数模型 半变系数模型 约束PLS估计 渐近正态性 

分 类 号：O212.7[理学—概率论与数理统计]