矩阵损失下有限总体中的局部线性Minimax预测

The Partial Linear Minimax Prediction in Finite Populations Under Matrix Loss Function

该文在矩阵损失下研究线性预测函数的局部极大极小性．在适当的假设下，得到了任意秩有限总体中的可预测变量的唯一的局部线性Ｍｉｎｉｍａｘ预测．（有关唯一性在几乎处处意义下理解）．

In this paper under the matrix loss function, the partial minimax property of linear prediction functions is investigated. Under suitable hypotheses, we obtain the unique partial linear minimax predictor of predictable variable in finite populations with arbitrary rank (we must comprehend uniqueness in the sense 'almost everywhere').