Premium
Learning Bayesian networks in the space of structures by estimation of distribution algorithms
Author(s) -
Blanco Rosa,
Inza Iñaki,
Larrañaga Pedro
Publication year - 2003
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.10084
Subject(s) - bayesian network , marginal likelihood , computer science , heuristics , population , univariate , estimation of distribution algorithm , bayesian probability , artificial intelligence , machine learning , bayesian information criterion , principle of maximum entropy , mathematics , algorithm , mathematical optimization , multivariate statistics , demography , sociology
The induction of the optimal Bayesian network structure is NP‐hard, justifying the use of search heuristics. Two novel population‐based stochastic search approaches, univariate marginal distribution algorithm (UMDA) and population‐based incremental learning (PBIL), are used to learn a Bayesian network structure from a database of cases in a score + search framework. A comparison with a genetic algorithm (GA) approach is performed using three different scores: penalized maximum likelihood, marginal likelihood, and information‐theory–based entropy. Experimental results show the interesting capabilities of both novel approaches with respect to the score value and the number of generations needed to converge. © 2003 Wiley Periodicals, Inc.