空间误差模型的变量选择

Variable selection for the spatial error model

基于惩罚拟似然的方法对空间误差模型的变量选择问题进行研究,得到了惩罚估计量的相合性和渐近正态性.通过蒙特卡洛模拟,验证了本文所提惩罚方法选择变量的有效性.

Based on penalized quasi-likelihood method,we studied the problem of variable selection in spatial error model,and obtained the consistency and asymptotic normality of penalty estimator.Monte Carlo simulation shows the effectiveness of the proposed penalty method for variable selection.