A Multi-parent Search Operator for Bayesian Network Building

David Iclnzan

Abstract. Learning a Bayesian network structure from data is a well-motivated but computationally hard task, especially for problems exhibiting synergic multivariate interactions. In this paper, a novel search method for structure learning of a Bayesian networks from binary data is proposed. The proposed method applies an entropy distillation operation over bounded groups of variables. A bias from the expected increase in randomness signals an underlaying statistical dependence between the inputs. The detected higher-order dependencies are used to connect linked attributes in the Bayesian network in a single step.

LNCS 7491, p. 246 ff.

Full article in PDF | BibTeX

lncs@springer.com
© Springer-Verlag Berlin Heidelberg 2012