The equivalence of Bayes and robust Bayes estimators for various loss functions

摘要

The problem of Bayes and robust Bayes estimation for various bounded and/or symmetric loss functions in a normalmodel with conjugate and non-informative prior distributions is considered. The prior distribution is not fully specified and covers the conjugate family of priors. It is of interest to know that the Bayes and robust Bayes estimators for symmetric losses are the same as those for the standard square-error loss function.