ORBi: Geurts Pierre - Investigation and reduction of discretization Variance in decision tree induction

Reference : Investigation and reduction of discretization Variance in decision tree induction


	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/2268/25762

Title :	Investigation and reduction of discretization Variance in decision tree induction
Language :	English
Author, co-author :	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 :	2000
Main document title :	Proceedings of ECML 2000, European Conference on Machine Learning
Publisher :	Springer-Verlag
Collection and collection volume :	LNAI 1810
Pages :	162-170
Peer reviewed :	Yes
On invitation :	No
Audience :	International
Event name :	European Conference on Machine Learning
Event date :	2000
Event place (city) :	Barcelona
Event country :	Spain
Keywords :	[en] machine learning
Abstract :	[en] This paper focuses on the variance introduced by the discretization techniques used to handle continuous attributes in decision tree induction. Different discretization procedures are first studied empirically, then means to reduce the discretization variance are proposed. The experiments shows that discretization variance is large and that it is possible to reduce it significantly without notable computational costs. The resulting variance reduction mainly improves interpretability and stability of decision trees, and marginally their accuracy.
Permalink :	http://hdl.handle.net/2268/25762
Other URL :	http://www.montefiore.ulg.ac.be/services/stochastic/pubs/2000/GW00

File(s) associated to this reference

Fulltext file(s):

	File	Commentary	Version	Size	Access
Open access	geurts-ecml2000.pdf		Publisher postprint	241.2 kB	View/Open

All documents in ORBi are protected by a user license.