半相依回归系统参数的Stein型主成分改进估计

The Stein-type Principal Components Adjusted Estimate of the Parameters in Seemingly Unrelated Regression System

将S tein压缩思想与主成分改进估计相结合,研究两个半相依回归系统.当设计阵呈病态时,提出S tein型主成分改进估计,同时,给出当V未知时的两步估计.证明在均方误差意义下,S tein型主成分改进估计局部优于主成分改进估计和协方差改进估计.

We would relate Stein-Type Shrunken thingking to the principal components adjusted estimate to study a seemingly unrelated regression system. A Stein-Type principal Components adjusted estimate is proposed as the design matrix is ill-conditioned while giving two-age estimate when V is unknown. Under the criteria of means square error,the Stein-Type principal components adjust estimate is better than the principal components adjust estimate and covariance adjusted estimate.