We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: An Empirical-Bayes Score for Discrete Bayesian Networks

Authors: Marco Scutari
Abstract: Bayesian network structure learning is often performed in a Bayesian setting, by evaluating candidate structures using their posterior probabilities for a given data set. Score-based algorithms then use those posterior probabilities as an objective function and return the maximum a posteriori network as the learned model. For discrete Bayesian networks, the canonical choice for a posterior score is the Bayesian Dirichlet equivalent uniform (BDeu) marginal likelihood with a uniform (U) graph prior (Heckerman et al., 1995). Its favourable theoretical properties descend from assuming a uniform prior both on the space of the network structures and on the space of the parameters of the network. In this paper, we revisit the limitations of these assumptions; and we introduce an alternative set of assumptions and the resulting score: the Bayesian Dirichlet sparse (BDs) empirical Bayes marginal likelihood with a marginal uniform (MU) graph prior. We evaluate its performance in an extensive simulation study, showing that MU+BDs is more accurate than U+BDeu both in learning the structure of the network and in predicting new observations, while not being computationally more complex to estimate.
Comments: 12 pages, PGM 2016
Subjects: Machine Learning (stat.ML); Methodology (stat.ME)
Journal reference: Journal of Machine Learning Research (52, Proceedings Track, PGM 2016), 438-448
Cite as: arXiv:1605.03884 [stat.ML]
  (or arXiv:1605.03884v3 [stat.ML] for this version)

Submission history

From: Marco Scutari [view email]
[v1] Thu, 12 May 2016 16:44:05 GMT (21kb)
[v2] Fri, 8 Jul 2016 11:24:40 GMT (21kb)
[v3] Mon, 13 Mar 2017 16:21:54 GMT (21kb)

Link back to: arXiv, form interface, contact.