矩阵损失下回归系数的条件MINIMAX估计

THE CONDITIONAL MINIMAX ESTIMATE OF REGRESSION COEFFICIENT UNDER MATRIX LOSS FUNCTION

考虑带线性等式约束的线性模型：ＥＹ＝ＸβＣｏｖ（Ｙ）＝σ２ＩＨβ＝０｛本文给出了矩阵损失下线性条件Ｍｉｎｉｍａｘ估计的定义；对任一条件可估函数Ｓβ。

Consider linear model with linear equality constraints EY=xβ\ \ Cov (y)=σ 2I Hβ=0 \ \ In this paper we define linear conditional minimax estimate under matrix loss function and obtain unique linear conditional minimax estimate for arbitrary conditional estimable function Sβ.