广义岭估计的谱分解式和Liu型广义岭估计

Spectral Decomposition Formula of Generalized Ridge Estimation and Liu-type Generalized Ridge Estimator

首先从多参数岭估计,即广义岭参数入手,给出其谱分解形式。然后就岭估计会使估值偏差增大的问题,在Liu估计的基础上,推导出一种新的有偏估计方法——Liu型广义岭估计,并给出该方法的模型、解式。Liu型广义岭估计可以在尽量不增大偏差的同时改进病态性,获得更好的估值。最后通过算例进行了比较分析。

This article presents the spectral decomposition formula of generalized ridge estimation. Then Liu-type estimator has been introduced. Based on it, Liu-type generalized ridge estimator has been put forward, which can improve the ill-conditioned equations without increasing the bias. By using a numerical example, some calculation and comparison are performed in the end.