Structural Learning of Bayesian Networks by Using Variable Neighbourhood Search Based on the Space of Orderings.
Comput. Syst. Dept., Univ. of Castilla-La Mancha, Albacete, Spain
DOI: 10.1109/ISDA.2009.157 Conference: Ninth International Conference on Intelligent Systems Design and Applications, ISDA 2009, Pisa, Italy , November 30-December 2, 2009
Structural learning of Bayesian networks (BNs) is an NP-hard problem generally addressed by means of heuristic search algorithms. Although these techniques do not guarantee an optimal result, they allow obtaining good solutions with a relatively low computational effort. Many proposals are based on searching the space of directed acyclic graphs. However, there are alternatives consisting of exploring the space of equivalence classes of BNs, which yields more complex and difficult to implement algorithms, or the space of the orderings among variables. In practice, ordering-based methods allow reaching good results, but, they are costly in terms of computation. In this paper, we prove the correctness of the method used to evaluate each permutation when exploring the space of orderings, and we propose two simple and efficient learning algorithms based on this approach. The first one is a Hill climbing method which uses an improved neighbourhood definition, whereas the second algorithm is its natural extension based on the well-known variable neighbourhood search metaheuristic. The algorithms have been tested over a set of different domains in order to study their behaviour in practice.
Available from: Jose Font
[Show abstract] [Hide abstract]
ABSTRACT: This paper proposes the EvoBANE system. EvoBANE automatically generates Bayesian networks for solving special-purpose problems.
EvoBANE evolves a population of individuals that codify Bayesian networks until it finds near optimal individual that solves
a given classification problem. EvoBANE has the flexibility to modify the constraints that condition the solution search space,
self-adapting to the specifications of the problem to be solved. The system extends the GGEAS architecture. GGEAS is a general-purpose
grammar-guided evolutionary automatic system, whose modular structure favors its application to the automatic construction
of intelligent systems. EvoBANE has been applied to two classification benchmark datasets belonging to different application
domains, and statistically compared with a genetic algorithm performing the same tasks. Results show that the proposed system
performed better, as it manages different complexity constraints in order to find the simplest solution that best solves every
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.