How large should ensembles of classifiers be?

This is the author’s version of a work that was accepted for publication in Pattern Recognition. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may ha...

Descripción completa

Detalles Bibliográficos
Autores: Hernández Lobato, Daniel, Martínez Muñoz, Gonzalo, Suárez González, Alberto
Tipo de recurso: artículo
Fecha de publicación:2013
País:España
Institución:Universidad Autónoma de Madrid
Repositorio:Biblos-e Archivo. Repositorio Institucional de la UAM
Idioma:inglés
OAI Identifier:oai:repositorio.uam.es:10486/664125
Acceso en línea:http://hdl.handle.net/10486/664125
https://dx.doi.org/10.1016/j.patcog.2012.10.021
Access Level:acceso abierto
Palabra clave:Asymptotic ensemble prediction
Bagging
Ensemble learning
Ensemble size
Random forest
Informática
id ES_062d6c59122273edd3ca66b4e3f1f5be
oai_identifier_str oai:repositorio.uam.es:10486/664125
network_acronym_str ES
network_name_str España
repository_id_str
spelling How large should ensembles of classifiers be?Hernández Lobato, DanielMartínez Muñoz, GonzaloSuárez González, AlbertoAsymptotic ensemble predictionBaggingEnsemble learningEnsemble sizeRandom forestInformáticaThis is the author’s version of a work that was accepted for publication in Pattern Recognition. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Pattern Recognition 46.5 (2013): 1323 – 1336, DOI: 10.1016/j.patcog.2012.10.021We propose to determine the size of a parallel ensemble by estimating the minimum number of classifiers that are required to obtain stable aggregate predictions. Assuming that majority voting is used, a statistical description of the convergence of the ensemble prediction to its asymptotic (infinite size) limit is given. The analysis of the voting process shows that for most test instances the ensemble prediction stabilizes after only a few classifiers are polled. By contrast, a small but non-negligible fraction of these instances require large numbers of classifier queries to reach stable predictions. Specifically, the fraction of instances whose stable predictions require more than T classifiers for T ≫ 1 has a universal form and is proportional to T−1/2. The ensemble size is determined as the minimum number of classifiers that are needed to estimate the infinite ensemble prediction at an average confidence level , close to one. This approach differs from previous proposals, which are based on determining the size for which the prediction error (not the predictions themselves) stabilizes. In particular, it does not require estimates of the generalization performance of the ensemble, which can be unreliable. It has general validity because it is based solely on the statistical description of the convergence of majority voting to its asymptotic limit. Extensive experiments using representative parallel ensembles (bagging and random forest) illustrate the application of the proposed framework in a wide range of classification problems. These experiments show that the optimal ensemble size is very sensitive to the particular classification problem considered.The authors acknowledge financial support from the Spanish Dirección General de Investigación, project TIN2010-21575-C02-02.Elsevier BVDepartamento de Ingeniería InformáticaEscuela Politécnica SuperiorAprendizaje Automático (ING EPS-001)20132013-05-01research articlehttp://purl.org/coar/resource_type/c_2df8fbb1AMhttp://purl.org/coar/version/c_ab4af688f83e57aainfo:eu-repo/semantics/articleapplication/pdfhttp://hdl.handle.net/10486/664125https://dx.doi.org/10.1016/j.patcog.2012.10.021reponame:Biblos-e Archivo. Repositorio Institucional de la UAMinstname:Universidad Autónoma de MadridInglésengopen accesshttp://purl.org/coar/access_right/c_abf2info:eu-repo/semantics/openAccessoai:repositorio.uam.es:10486/6641252026-06-23T12:46:27Z
dc.title.none.fl_str_mv How large should ensembles of classifiers be?
title How large should ensembles of classifiers be?
spellingShingle How large should ensembles of classifiers be?
Hernández Lobato, Daniel
Asymptotic ensemble prediction
Bagging
Ensemble learning
Ensemble size
Random forest
Informática
title_short How large should ensembles of classifiers be?
title_full How large should ensembles of classifiers be?
title_fullStr How large should ensembles of classifiers be?
title_full_unstemmed How large should ensembles of classifiers be?
title_sort How large should ensembles of classifiers be?
dc.creator.none.fl_str_mv Hernández Lobato, Daniel
Martínez Muñoz, Gonzalo
Suárez González, Alberto
author Hernández Lobato, Daniel
author_facet Hernández Lobato, Daniel
Martínez Muñoz, Gonzalo
Suárez González, Alberto
author_role author
author2 Martínez Muñoz, Gonzalo
Suárez González, Alberto
author2_role author
author
dc.contributor.none.fl_str_mv Departamento de Ingeniería Informática
Escuela Politécnica Superior
Aprendizaje Automático (ING EPS-001)
dc.subject.none.fl_str_mv Asymptotic ensemble prediction
Bagging
Ensemble learning
Ensemble size
Random forest
Informática
topic Asymptotic ensemble prediction
Bagging
Ensemble learning
Ensemble size
Random forest
Informática
description This is the author’s version of a work that was accepted for publication in Pattern Recognition. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Pattern Recognition 46.5 (2013): 1323 – 1336, DOI: 10.1016/j.patcog.2012.10.021
publishDate 2013
dc.date.none.fl_str_mv 2013
2013-05-01
dc.type.none.fl_str_mv research article
http://purl.org/coar/resource_type/c_2df8fbb1
AM
http://purl.org/coar/version/c_ab4af688f83e57aa
dc.type.openaire.fl_str_mv info:eu-repo/semantics/article
format article
dc.identifier.none.fl_str_mv http://hdl.handle.net/10486/664125
https://dx.doi.org/10.1016/j.patcog.2012.10.021
url http://hdl.handle.net/10486/664125
https://dx.doi.org/10.1016/j.patcog.2012.10.021
dc.language.none.fl_str_mv Inglés
eng
language_invalid_str_mv Inglés
language eng
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Elsevier BV
publisher.none.fl_str_mv Elsevier BV
dc.source.none.fl_str_mv reponame:Biblos-e Archivo. Repositorio Institucional de la UAM
instname:Universidad Autónoma de Madrid
instname_str Universidad Autónoma de Madrid
reponame_str Biblos-e Archivo. Repositorio Institucional de la UAM
collection Biblos-e Archivo. Repositorio Institucional de la UAM
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869402900425342976
score 15.301603