MINIMA OF CONVEX INTEGRAL FUNCTIONALS AND UNBIASED
ESTIMATION
Abstract: Necessary and sufficient conditions for the optimality of unbiased estimators in case of
arbitrary finite convex loss functions are given. These conditions are derived from a theorem
on subdif-ferentials of convex integral functionals on Orlicz spaces. The results
obtained provide a basic tool for problems, concerning universal loss functions and
considered in paper [12]. They are also related to Cramér-Rao type inequalities.
2000 AMS Mathematics Subject Classification: Primary: -; Secondary: -;
Key words and phrases: -