岭回归估计β(k)的一个特性及其应用

A PROPERTY OF RIDGE REGRESSION ESTIMAPER β(k) AND ITS' APPLICATION

本文就岭回归估计β_(k)在0<K≤σ~2/maxr_i^2条件下,对标准E‖β_(k)-β‖~2优于LS估计问题给出了证明,并结合实例确定较优的岭回归K值。

In this paper we prove that ridge regression estimater ?(K) under the condition of 0<K≤σ2/maxr;2 is more superier to the LS estimater of standard E||β(K)-β||2 .We also give examples for showing the superier ridge regression K-Value.