摘要
从矩阵变换理论出发,对多元线性模型的系数提出了广义根方估计。证明了它优于系数的LS估计,根方改进估计,且是β的线性可容性估计等性质。
A new generalized root-root estimator is considered in multivariate linear model.By using the matrix transformation theory,it proved the generalized root-root estimator,under the quadratic loss function,is superior to both LSE and RRE,and is also a linear admissible estimator of the regression coefficient.
出处
《西北大学学报(自然科学版)》
CAS
CSCD
1996年第6期467-470,共4页
Journal of Northwest University(Natural Science Edition)
关键词
多元线性模型
可容许估计
广义根方估计
multivariate linear model
generalized root-root estimator
admissible estimator
mean square error