Original title: On maximization of the information divergence from an exponential family
Authors: Matúš, František ; Ay, N.
Document type: Papers
Conference/Event: WUPES 2003. Workshop on Uncertainty Processing /6./, Hejnice (CZ), 2003-09-24 / 2003-09-27
Year: 2003
Language: eng
Abstract: The information divergence of a probability measure P from an exponential family E over a finite set is defined as infimum of the divergences of P from Q subject to Q in E. For convex exponential families the local maximizers of this function of P are found. General exponential family E of dimension d is enlarged to an exponential family E* of the dimension at most 3d+2 such that the local maximizers are of zero divergence from E*.
Keywords: exponential family; information projection; Kullback-Leibler divergence
Project no.: CEZ:AV0Z1075907 (CEP), IAA1075104 (CEP), GA402/01/0981 (CEP)
Funding provider: GA AV ČR, GA ČR
Host item entry: Proceedings of the 6th Workshop on Uncertainty Processing

Institution: Institute of Information Theory and Automation AS ČR (web)
Document availability information: Fulltext is available at the institute of the Academy of Sciences.
Original record: http://hdl.handle.net/11104/0131224

Permalink: http://www.nusl.cz/ntk/nusl-34998


The record appears in these collections:
Research > Institutes ASCR > Institute of Information Theory and Automation
Conference materials > Papers