带有误差线性协变量的半参数变系数部分线性模型的协变量选择(英文)

Covariate Selection for Semiparametric Varying-Coefficient Partially Linear Models with Error-Prone Linear Covariates

考虑了一些协变量不被观察到,但辅助变量可用时的半参数变系数部分线性模型.提出了基于SCAD惩罚的参数分量的变量选择方法,且估计既有渐近正态性又有先知性.模拟研究说明了所提出变量选择方法的有限样本性质.

This paper considers semiparametric varying-coefficient partially linear models when some eovariates are not observed, but ancillary variables are available. A variable selection procedure for the parametric components is presented based on the SCAD penalization, and the resulting estimates perform asymptotic normality as well as an oracle property. Finite sample performance of the proposed variable selection procedures is illustrated by simulation studies.