桥回归的系数估计对应于gamma-范数的压缩规律
General Shrinkage Rule with Respect to gamma-Norms for Bridge Estimation
-
摘要: 惩罚函数为, 的桥回归, 作为一类特殊的惩罚回归方法, 已经被很多文献研究过. 本文分别给出了估计系数的压缩大小与在和两种不同情况下取值的一些理论结果, 并通过模拟来验证压缩估计的估计效果.Abstract: Bridge regression, a special family of penalized regressions of a penalty function with , has been studied in many literatures. In this paper, we provide some theoretical results of how the shrinkage rule changing with under two settings: and , respectively. Simulation results are conducted to evaluate the performance of the proposed method.