不变二次损失下期望向量的线性容许估计

Linear Admissible Estimators of Mean Vector under Invariant Quadratic Loss

摘要: 设Y1，Y2，…，Y_n独立同分布，EY₁＝β,CovY₁＝∑,β∈R^p和∑＞0均未知。在不变损失函数（d－β）＇∑^-1（d－β）下，我们给出了齐次和非齐次线性估计在各自的类中是β的容许估计的充要条件。

Abstract: Let Y1，Y2，…，Y_n, be i.i.d., EY₁＝β,CovY₁＝∑,β∈R^p and ∑＞0 are unknown parameters.Under invariant quadratic loss function, we obtained necessary and sufficient condition that a linear estimator is admissible among linear estimator class.