部分线性模型的模态正交经验似然推断

本文考虑部分线性模型的有效经验似然统计推断问题.通过结合模态回归和正交投影技术,提出了一种模态经验似然统计推断过程.证明了提出的经验似然比函数渐近服从中心卡方分布,进而构造了模型参数的置信区间.所提出的估计方法可以对模型的参数分量和非参数分量分别估计,而互不影响,具有较好的稳健性和有效性.

Orthogonality Based Modal Empirical Likelihood for Partially Linear Models

This paper considers the effective empirical likelihood inference for partially linear models.By combining modal regression method with orthogonal projection technology,a modal empirical likelihood based estimation procedure is proposed.Under some mild conditions,we show that Wilks’theorem of the proposed empirical likelihood approach continues to hold,and then the confidence regions of model coefficients are constructed.The resulting estimators for parametric and nonparametric components do not affect each other,and then the proposed method is more robust and effective.