部分线性可加模型基于backfitting方法的岭估计

Backfitting-based Ridge Estimation in Semiparametric Patially Linear Additive Models

作为部分线性模型和可加模型的推广,部分线性可加模型是一类应用广泛的半参数模型.本文主要考虑该模型在线性部分自变量存在共线性情形下的估计问题.首先我们构造了参数分量的岭估计,其次我们构造了含有约束条件时的岭估计.

As an extention of partially linear additive model and additive models,Partially linear additive model is widely used in semi-parametric models.This paper focuses on the case that there is multicollinearity exists among independent variables of linear part of model.Firstly,we construct a class of ridge estimators for the parametric component.When some additional linear restrictions on the parametric component are available,we proposed restricted ridge estimators based on the backfitting approaches too.