Under multivariant normal linear models, it is proved that usual loss estima-tors based on the uniformly minimum variance unbiased estimator and the best affinelyequivariant estimator of error variance are inadmissibl...Under multivariant normal linear models, it is proved that usual loss estima-tors based on the uniformly minimum variance unbiased estimator and the best affinelyequivariant estimator of error variance are inadmissible with squared error loss.展开更多
Convergence rates of empirical Bayes(EB) estimators w-r-t the squared efror loss are con sidered in discrete exponential families. It is shown that a rate O(n^(-1)) is not possible for any EB estimator,even though the...Convergence rates of empirical Bayes(EB) estimators w-r-t the squared efror loss are con sidered in discrete exponential families. It is shown that a rate O(n^(-1)) is not possible for any EB estimator,even though the parameter space is bounded. Namely, it is proved that the discrete similar of one conjecture of Singh (for Lebesgue-exponential families) is correct.展开更多
A theorem concerning a conjecture of Singh is formulated in[2].But the argumentin [2] contains a serious gap which is in fact the essential point of the proof.A correct proofis presented here.
基金Supported by National Natural Science Foundation of China.
文摘Under multivariant normal linear models, it is proved that usual loss estima-tors based on the uniformly minimum variance unbiased estimator and the best affinelyequivariant estimator of error variance are inadmissible with squared error loss.
文摘Convergence rates of empirical Bayes(EB) estimators w-r-t the squared efror loss are con sidered in discrete exponential families. It is shown that a rate O(n^(-1)) is not possible for any EB estimator,even though the parameter space is bounded. Namely, it is proved that the discrete similar of one conjecture of Singh (for Lebesgue-exponential families) is correct.
文摘A theorem concerning a conjecture of Singh is formulated in[2].But the argumentin [2] contains a serious gap which is in fact the essential point of the proof.A correct proofis presented here.
