ORBi: Schnitzler François - Efficiently approximating Markov tree bagging for high-dimensional density estimation

Reference : Efficiently approximating Markov tree bagging for high-dimensional density estimation


	Document type :	Scientific congresses and symposiums : Paper published in a book
	Discipline(s) :	Engineering, computing & technology : Computer science
	To cite this reference:	http://hdl.handle.net/2268/96741

Title :	Efficiently approximating Markov tree bagging for high-dimensional density estimation
Language :	English
Author, co-author :	Schnitzler, François [Université de Liège - ULg > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
	ammar, sourour [ > > ]
	leray, philippe [ > > ]
	Geurts, Pierre [Université de Liège - ULg > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
	Wehenkel, Louis [Université de Liège - ULg > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
Publication date :	Sep-2011
Main document title :	Machine Learning and Knowledge Discovery in Databases, Part III
Editor :	Gunopulos, Dimitrios
	Hofmann, Thomas
	Malerba, Donato
	Vazirgiannis, Michalis
Publisher :	Springer-Verlag
Collection and collection volume :	LNAI 6913
Pages :	113-128
On invitation :	No
Audience :	International
ISBN :	978-3-642-23807-9
City :	Berlin, Heidelberg
Event name :	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Event date :	from 05-09-2011 to 09-09-2011
Event organizer :	Prof. A. Likas
	Prof. Y. Theodoridis
Event place (city) :	Athens
Event country :	Greece
Keywords :	[en] bayesian networks ; Markov trees ; mixture of trees ; bagging ; bootstrap ; Chow-Liu algorithm
Abstract :	[en] We consider algorithms for generating Mixtures of Bagged Markov Trees, for density estimation. In problems deﬁned over many variables and when few observations are available, those mixtures generally outperform a single Markov tree maximizing the data likelihood, but are far more expensive to compute. In this paper, we describe new algorithms for approximating such models, with the aim of speeding up learning without sacriﬁcing accuracy. More speciﬁcally, we propose to use a ﬁltering step obtained as a by-product from computing a ﬁrst Markov tree, so as to avoid considering poor candidate edges in the subsequently generated trees. We compare these algorithms (on synthetic data sets) to Mixtures of Bagged Markov Trees, as well as to a single Markov tree derived by the classical Chow-Liu algorithm and to a recently proposed randomized scheme used for building tree mixtures.
Research units :	Systèmes et Modélisation
Funders :	Fonds pour la formation à la Recherche dans l'Industrie et dans l'Agriculture (Communauté française de Belgique) - FRIA ; Biomagnet IUAP network of the Belgian Science Policy Office ; Pascal2 network of excellence of the EC
Target :	Researchers
Permalink :	http://hdl.handle.net/2268/96741
DOI :	10.1007/978-3-642-23808-6_8
Other URL :	http://www.springerlink.com/content/97100k1308156219/
Mentions required by the publisher for OA :	The original publication is available at www.springerlink.com

File(s) associated to this reference

Fulltext file(s):

	File	Commentary	Version	Size	Access
Open access	schnitzler_363.pdf		Author postprint	222.81 kB	View/Open

Additional material(s):

	File	Commentary	Version	Size	Access
Open access	ECML_poster.pdf	poster	-	321.92 kB	View/Open
Open access	ECML_presentation.pdf	Conference slides	-	284.81 kB	View/Open

All documents in ORBi are protected by a user license.

Dspace v1.5