Relevance Measures for Localized Partial Evaluation of Belief Networks

Denise Draper

Localized partial evaluation is an algorithm for computing bounds on the marginal probability of a variable in a belief network. LPE accomplishes this by considering information incrementally, attempting to find more relevant information first. In this paper, we briefly describe belief networks, the localized partial evaluation algorithm, and then discuss how relevance can be defined and used in LPE.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.