Approximate structure learning for large Bayesian networks

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

Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

25 Citaten (Scopus)
69 Downloads (Pure)

Samenvatting

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.

Originele taal-2Engels
Pagina's (van-tot)1209-1227
Aantal pagina's19
TijdschriftMachine Learning
Volume107
Nummer van het tijdschrift8-10
DOI's
StatusGepubliceerd - 1 sep. 2018
Extern gepubliceerdJa

Vingerafdruk

Duik in de onderzoeksthema's van 'Approximate structure learning for large Bayesian networks'. Samen vormen ze een unieke vingerafdruk.

Citeer dit