部分线性可加模型的约束估计

Restricted Estimation of Partially Linear Additive Models

文章讨论部分线性模型在约束情况下的估计问题.在一般的线性约束条件下利用两种方法构造参数分量的约束估计;当参数部分附加随机约束条件时,依据混合估计方法与Profile最小二乘方法,构造参数分量的随机约束估计;讨论一般约束估计和随机约束估计的关系.

This paper discusses statistical inference for the semiparametric model when some additional restrictions on the parametric component are available. We first propose a restricted estimator for the regression coefficients by two ways respectively. Then, when there are stochastic linear restrictions on the parametric component, we introduce a new estimator by combining ideas underlying the mixed estimator and the profile least- squares approach. Finally, the relation between the restricted estimator and the stochastic restricted estimator is discussed.