摘要
对于相依线性回归系统(SUR,m=2),其设计阵为病态时,文中给出了βi(i=1,2)的压缩估计及相应的两步估计,证明了在θi(i=1,2,…,p1)满足一定条件时,此估计相对于协方差改进估计及两步协方差改进估计的优良性,同时给出了θi(i=1,2,…,p1)的确定方法.
For seemingly unrelated regression equations with design matrices being ill-conditioned,we proposed the shrunken principal estimators and the related two-stage estimators of βi(i=1,2).When θi(i=1,2,…,p1) satisfied a certain condition,we proved the superiority to CI estimator and two-stage CI estimator.Finally,we gave a method to find the solution of θi(i=1,2,…,p1).
出处
《江苏科技大学学报(自然科学版)》
CAS
北大核心
2010年第2期197-200,共4页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
基金
江苏省高校自然科学基金资助项目(09KJD110002)
关键词
相依线性回归系统
压缩估计
协方差改进估计
两步估计
seemingly unrelated regression model
shrunken principal estimator
covariance improvement estimator
two-stage estimator