A Caltech Library Service

Maximum Entropy Relaxation for Graphical Model Selection given Inconsistent Statistics

Chandrasekaran, Venkat and Johnson, Jason K. and Willsky, Alan S. (2007) Maximum Entropy Relaxation for Graphical Model Selection given Inconsistent Statistics. In: IEEE/SP 14th Workshop on Statistical Signal Processing. IEEE , Piscataway, NJ, pp. 625-629. ISBN 978-1-4244-1197-9.

PDF - Submitted Version
See Usage Policy.

PDF - Published Version
See Usage Policy.


Use this Persistent URL to link to this item:


We develop a novel approach to approximate a specified collection of marginal distributions on subsets of variables by a globally consistent distribution on the entire collection of variables. In general, the specified marginal distributions may be inconsistent on overlapping subsets of variables. Our method is based on maximizing entropy over an exponential family of graphical models, subject to divergence constraints on small subsets of variables that enforce closeness to the specified marginals. The resulting optimization problem is convex, and can be solved efficiently using a primal-dual interiorpoint algorithm. Moreover, this framework leads naturally to a solution that is a sparse graphical model.

Item Type:Book Section
Related URLs:
URLURL TypeDescription
Additional Information:© 2007 IEEE. Date of Current Version: 17 September 2007.
Subject Keywords:graphical models; maximum entropy principle; model selection; inconsistent statistics
Record Number:CaltechAUTHORS:20121009-080604998
Persistent URL:
Official Citation:Chandrasekaran, Venkat; Johnson, Jason K.; Willsky, Alan S.; , "Maximum Entropy Relaxation for Graphical Model Selection Given Inconsistent Statistics," Statistical Signal Processing, 2007. SSP '07. IEEE/SP 14th Workshop on , vol., no., pp.625-629, 26-29 Aug. 2007 doi: 10.1109/SSP.2007.4301334 URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:34769
Deposited By: Tony Diaz
Deposited On:09 Oct 2012 20:08
Last Modified:09 Nov 2021 23:10

Repository Staff Only: item control page