部分线性变系数模型的Liu估计

Liu Estimation in Partially Linear Varying Coefficient Models

研究了部分线性变系数模型中参数分量的有偏估计问题.基于Profile最小二乘方法和Liu估计法构造了参数分量的Profile-Liu估计和剖面最小二乘广义Liu估计,并在一定的条件下,证明了Profile-Liu估计优于Profile最小二乘估计,剖面最小二乘广义Liu估计优于Profile-Liu估计.

This paper considers the biased estimators on the parameter components in partially linear varying coefficient model.Based on the Profile least-square approach and Liu estimation method, we propose Profile-Liu estimator and section least-square general Liu estimator for the parametric components.We prove that the Profile-Liu estimation is better than the Profile least-square estimation, and the section least-square general Liu estimation is better than the Profile-Liu estimation under certain conditions.