Approximate structure learning for large Bayesian networks

Mauro Scanagatta, Giorgio Corani, Cassio Polpo de Campos, Marco Zaffalon

Research output: Contribution to journalArticleAcademicpeer-review

13 Citations (Scopus)
9 Downloads (Pure)

Abstract

We present approximate structure learning algorithms for Bayesian networks. We discuss the two main phases of the task: the preparation of the cache of the scores and structure optimization, both with bounded and unbounded treewidth. We improve on state-of-the-art methods that rely on an ordering-based search by sampling more effectively the space of the orders. This allows for a remarkable improvement in learning Bayesian networks from thousands of variables. We also present a thorough study of the accuracy and the running time of inference, comparing bounded-treewidth and unbounded-treewidth models.

Original languageEnglish
Pages (from-to)1209-1227
Number of pages19
JournalMachine Learning
Volume107
Issue number8-10
DOIs
Publication statusPublished - 1 Sep 2018
Externally publishedYes

Keywords

  • Bayesian networks
  • Structural learning
  • Treewidth

Fingerprint Dive into the research topics of 'Approximate structure learning for large Bayesian networks'. Together they form a unique fingerprint.

Cite this