Abstract : In this paper, we consider a Bayesian network structure estimation problem as a two step problem based on an independence model representation. We first perform an evolutionary search for an approximation of an independence model. A deterministic algorithm is then used to deduce a Bayesian network which represents the equivalence class of the independence model. This paper is a shortened version of a paper that has been published in a genetic algorithms conference [2].
https://hal.archives-ouvertes.fr/hal-00467586
Contributeur : Evelyne Lutton <>
Soumis le : mardi 4 mai 2010 - 17:31:40
Dernière modification le : jeudi 9 février 2017 - 15:48:08
Document(s) archivé(s) le : mercredi 30 novembre 2016 - 16:51:05
Olivier Barrière, Evelyne Lutton, Pierre-Henri Wuillemin. Independence Model estimation using Artificial Evolution. 5èmes Journées Francophones sur les Réseaux Bayésiens (JFRB2010), May 2010, Nantes, France. <hal-00467586>
