Title :
Characterization of the Bayes estimator and the MDL estimator for exponential families
Author :
Takeuchi, Jun-ichi
Author_Institution :
NEC Corp., Japan
Abstract :
We analyze the relationship between the MDL (minimum description length) estimator (posterior mode) and the PBE (projected Bayes estimator) for exponential families, where the PBE is obtained by projecting the BE (Bayes estimator, i.e. posterior mixture) onto the original exponential family and is equal to the BE under a certain condition
Keywords :
Bayes methods; estimation theory; parameter estimation; Bayes estimator; MDL estimator; exponential families; minimum description length; parameter estimation; posterior mixture; posterior mode; projected Bayes estimator; Information geometry; Laboratories; Maximum likelihood estimation; Minimax techniques; National electric code; Random variables;
Conference_Titel :
Information Theory, 1995. Proceedings., 1995 IEEE International Symposium on
Conference_Location :
Whistler, BC
Print_ISBN :
0-7803-2453-6
DOI :
10.1109/ISIT.1995.535743
