ORBi: Joly Arnaud - L1-based compression of random forest models

Reference : L1-based compression of random forest models


	Document type :	Scientific congresses and symposiums : Paper published in a book
	Discipline(s) :	Engineering, computing & technology : Computer science Engineering, computing & technology : Electrical & electronics engineering
	To cite this reference:	http://hdl.handle.net/2268/112824

Title :	L1-based compression of random forest models
Language :	English
Author, co-author :	Joly, Arnaud [Université de Liège - ULg > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
	Schnitzler, François [Université de Liège - ULg > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
	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 :	Apr-2012
Main document title :	20th European Symposium on Artificial Neural Networks
On invitation :	No
Audience :	International
Event name :	European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Event date :	25 - 27 April 2012
Event organizer :	Michel Verleysen
Event place (city) :	Bruges
Event country :	Belgium
Keywords :	[en] Ensemble of randomized trees ; Pruning ; L1-norm regularization ; LASSO ; Supervised learning ; Machine Learning ; Randomization ; Model reduction ; Decision tree
Abstract :	[en] Random forests are effective supervised learning methods applicable to large-scale datasets. However, the space complexity of tree ensembles, in terms of their total number of nodes, is often prohibitive, specially in the context of problems with very high-dimensional input spaces. We propose to study their compressibility by applying a L1-based regularization to the set of indicator functions defined by all their nodes. We show experimentally that preserving or even improving the model accuracy while significantly reducing its space complexity is indeed possible.
Research units :	Système et modélisation ; GIGA-research
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 ; Professionals
Permalink :	http://hdl.handle.net/2268/112824

File(s) associated to this reference

Fulltext file(s):

	File	Commentary	Version	Size	Access
Open access	es2012-43.pdf		Publisher postprint	242.77 kB	View/Open

All documents in ORBi are protected by a user license.

Dspace v1.5