摘要
考虑纵向数据下的变系数回归模型y（ij）=xij^Tθ（t（ij）＋eiji=1,2,…,n j=1,2,…,m.利用小波光滑和加权最小二乘方法,分别研究了模型中未知参数θ（·）的小波估计θ（·）和误差方差σ^2的小波估计σ^2,在适当的条件下,证明了θ的强相合性,强相合速度,并得到θ和σ^2的渐近正态性.
In this article, varying-coefficient regression models with longitudinal data are investigated. Combining with wavelet smoothing and weighted least squares, we develop wavelet estimators of unknown parameter and error variance, and under mild conditions, their asymptotic normalities, strong consistency and convergence rates are given.
出处
《数学的实践与认识》
北大核心
2015年第8期271-278,共8页
Mathematics in Practice and Theory
关键词
纵向数据
变系数模型
小波估计
渐近正态性
longitudinal data
varying-coefficient regression models
wavelet estimators asymptotic normality
convergence rates
