增长曲线模型中系数矩阵的线性容许Minimax估计

The Linear Admissible Minimax Estimate of Regression Coefficient in Growth Curve Model

对于生长曲线模型Y＝X_1B_1X_2＋ε，Cov（ε）＝σV，本文分别在某种齐次线性估计类L0和非齐次线性估计类L_1中找到了系数矩阵的线性可估函数KBL的容许Minimax估计，并且证明了这种估计是唯一的．

Given a growth curve model EY= X_1BX_2 + ε,Cov(ε) =σ~2V, the estimable function KBL is to be required to be estimated, where K, L are resperately m1 X k and q X m_2 constant matrix. In this paper under matrix loss functions the admissible minimax property of the estimate in the class of homogenous and no-homogeneous linear estimate is studied.