摘要
Zellner等学者研究了一类相依回归系统中参数的最小二乘估计的改进,这些改进估计都是无编估计且具有许多优良性质,但是当设计阵呈病态时这些改进估计不再被认为是良好估计.本文提出了一种有偏Stein估计,证明了在设计阵呈病态时Stein估计的优良性质.
One kind of least square estimators of the parameters in two seemingly unrelated regression systems have been improved by Zellner,et al(1).These improved estimatora are all unbiased estimators with a lot of advantages,however,they are no longer considered to be well estimators in case the design matrix is ill-conditioning.In this paper,the author proposes one kind of biased Stein estimators;and demonstrates the advantages of Stein estimatiors in case the design matrix is ill-conditioning.
出处
《华侨大学学报(自然科学版)》
CAS
1993年第2期154-158,共5页
Journal of Huaqiao University(Natural Science)
基金
国务院侨办基金资助课题
关键词
半相依回归
斯坦因估计
无偏估计
semingly unrelated regression systems
two step estimate
covariance improved
estimation