摘要
研究回归系数岭型主相关估计在降维估计类中的方差最优性,证明它的方差阵最小,方差阵的特征值最小,并讨论它的方差在正交不变范数意义下的最小性及方差的最值性。
In this paper the variance optimality of combining ridge and principal correlation estimation is discussed in the class of reduced-dimension estimates. And it is proved that it has minimum variance of sum and multiplication of variance. The maximum property of variance in the sense of orthogonal unchanged norm is studied.
出处
《廊坊师范学院学报（自然科学版）》
2009年第4期14-15,共2页
Journal of Langfang Normal University(Natural Science Edition)
关键词
回归模型
岭型主相关估计
方差
最优性质
linear regression model
combining ridge and principal correlation estimation
optimal property
