Inizio della pagina -
Logo DISCO
|
Visita la Versione ad elevata leggibilità
|
Vai al Contenuto della pagina
|
Vai alla Fine dei contenuti
|
Vai al Menu Principale
|
Vai alla Barra di navigazione (sei in)
|
Vai al Menu di navigazione (albero)
|
Vai alla Lista dei comandi
|
Vai alla Lista degli approfondimenti
|
Vai al Menu inferiore
|
Logo Ateneo
   
An Empirical-Bayes Score for Discrete Bayesian Networks

17 Gennaio 2017, ore 11:30

Sala Seminari, 1° piano - Dipartimento di Informatica, Sistemistica e Comunicazione, Edificio U14 
Relatore/i: Dr. 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.

 

 

Short bio:

Marco studied Statistics and Computer Science at the University of Padova, Italy. He earned his Ph.D. in Statistics in Padova under the guidance of Professor A. Brogini, with a thesis on graphical modelling. He then moved to University College London (UCL) as a Research Associate in Statistical Genetics at the Genetics Institute (UGI). His research focuses on the theory of Bayesian networks and their applications to biological data, and he is the author and maintainer of the bnlearn R package.

 

For information contact Prof. Fabio Stella stella@disco.unimib.it

In archivio dal: 18/01/2017


  

  

Google Translate
Translate to English Translate to French Translate to German Translate to Spanish Translate to Chinese Translate to Portuguese Translate to Arabic
Translate to Albanian Translate to Bulgarian Translate to Croatian Translate to Czech Translate to Danish Translate to Dutch Translate to Finnish Translate to Greek Translate to Hindi
Translate to Hungarian Translate to Irish Translate to Japanese Translate to Korean Translate to Norwegian Translate to Polish Translate to Romanian Translate to Russian Translate to Serbian
Translate to Slovenian Translate to Swedish Translate to Thai Translate to Turkish

(C) Copyright 2016 - Dipartimento Informatica Sistemistica e Comunicazione - Viale Sarca, 336
20126 Milano - Edificio U14
redazioneweb@disco.unimib.it - ultimo aggiornamento di questa pagina 13/09/2017