Title :
On maximum likelihood estimation and divergence minimization
Author :
Chan, Terence H. ; Yeung, Raymond W.
Author_Institution :
Dept. of Electr. & Comput. Eng., Toronto Univ., Ont., Canada
Abstract :
We reformulate probabilistic inference problems as divergence minimization (DM) problems, which have a natural geometrical interpretation. An iterative algorithm is proposed to solve the DM problem. It turns out that the EM algorithm could be regarded as a special case.
Keywords :
inference mechanisms; iterative methods; maximum likelihood estimation; minimisation; probability; EM algorithm; MLE; divergence minimization; divergence minimization problems; geometrical interpretation; iterative algorithm; maximum likelihood estimation; probabilistic inference problems; Delta modulation; Inference algorithms; Information geometry; Iterative algorithms; Maximum likelihood estimation; Minimization methods; Parametric statistics; Random variables; Solid modeling;
Conference_Titel :
Information Theory, 2002. Proceedings. 2002 IEEE International Symposium on
Print_ISBN :
0-7803-7501-7
DOI :
10.1109/ISIT.2002.1023430
