A mathematical programming approach to SVM-based classification with label noise

In this paper we propose novel methodologies to optimally construct Support Vector Machine-based classifiers that take into account that label noise occur in the training sample. We propose different alternatives based on solving Mixed Integer Linear and Non Linear models by incorporating decisions...

ver descrição completa

Detalhes bibliográficos
Autores: Blanco, Víctor, Japón Sáez, Alberto, Puerto Albandoz, Justo
Formato: artículo
Estado:Versión publicada
Fecha de publicación:2022
País:España
Recursos:Universidad de Sevilla (US)
Repositorio:idUS. Depósito de Investigación de la Universidad de Sevilla
OAI Identifier:oai:idus.us.es:11441/144676
Acesso em linha:https://hdl.handle.net/11441/144676
https://doi.org/10.1016/j.cie.2022.108611
Access Level:acceso abierto
Palavra-chave:Supervised classification
SVM
Mixed integer non linear programming
Label noise
id ES_7b239e6c082c4b6e0d7ffb4fd95111ef
oai_identifier_str oai:idus.us.es:11441/144676
network_acronym_str ES
network_name_str España
repository_id_str
spelling A mathematical programming approach to SVM-based classification with label noiseBlanco, VíctorJapón Sáez, AlbertoPuerto Albandoz, JustoSupervised classificationSVMMixed integer non linear programmingLabel noiseIn this paper we propose novel methodologies to optimally construct Support Vector Machine-based classifiers that take into account that label noise occur in the training sample. We propose different alternatives based on solving Mixed Integer Linear and Non Linear models by incorporating decisions on relabeling some of the observations in the training dataset. The first method incorporates relabeling directly in the SVM model while a second family of methods combines clustering with classification at the same time, giving rise to a model that applies simultaneously similarity measures and SVM. Extensive computational experiments are reported based on a battery of standard datasets taken from UCI Machine Learning repository, showing the effectiveness of the proposed approaches.ScienceDirectEstadística e Investigación OperativaFQM331: Metodos y Modelos de la Estadistica y la Investigacion Operativa2022info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfapplication/pdfhttps://hdl.handle.net/11441/144676https://doi.org/10.1016/j.cie.2022.108611reponame:idUS. Depósito de Investigación de la Universidad de Sevillainstname:Universidad de Sevilla (US)InglésCOMPUTERS & INDUSTRIAL ENGINEERING, 172 (Parte A), 108611-1.https://doi.org/10.1016/j.cie.2022.108611info:eu-repo/semantics/openAccessoai:idus.us.es:11441/1446762026-06-17T12:51:07Z
dc.title.none.fl_str_mv A mathematical programming approach to SVM-based classification with label noise
title A mathematical programming approach to SVM-based classification with label noise
spellingShingle A mathematical programming approach to SVM-based classification with label noise
Blanco, Víctor
Supervised classification
SVM
Mixed integer non linear programming
Label noise
title_short A mathematical programming approach to SVM-based classification with label noise
title_full A mathematical programming approach to SVM-based classification with label noise
title_fullStr A mathematical programming approach to SVM-based classification with label noise
title_full_unstemmed A mathematical programming approach to SVM-based classification with label noise
title_sort A mathematical programming approach to SVM-based classification with label noise
dc.creator.none.fl_str_mv Blanco, Víctor
Japón Sáez, Alberto
Puerto Albandoz, Justo
author Blanco, Víctor
author_facet Blanco, Víctor
Japón Sáez, Alberto
Puerto Albandoz, Justo
author_role author
author2 Japón Sáez, Alberto
Puerto Albandoz, Justo
author2_role author
author
dc.contributor.none.fl_str_mv Estadística e Investigación Operativa
FQM331: Metodos y Modelos de la Estadistica y la Investigacion Operativa
dc.subject.none.fl_str_mv Supervised classification
SVM
Mixed integer non linear programming
Label noise
topic Supervised classification
SVM
Mixed integer non linear programming
Label noise
description In this paper we propose novel methodologies to optimally construct Support Vector Machine-based classifiers that take into account that label noise occur in the training sample. We propose different alternatives based on solving Mixed Integer Linear and Non Linear models by incorporating decisions on relabeling some of the observations in the training dataset. The first method incorporates relabeling directly in the SVM model while a second family of methods combines clustering with classification at the same time, giving rise to a model that applies simultaneously similarity measures and SVM. Extensive computational experiments are reported based on a battery of standard datasets taken from UCI Machine Learning repository, showing the effectiveness of the proposed approaches.
publishDate 2022
dc.date.none.fl_str_mv 2022
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv https://hdl.handle.net/11441/144676
https://doi.org/10.1016/j.cie.2022.108611
url https://hdl.handle.net/11441/144676
https://doi.org/10.1016/j.cie.2022.108611
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.relation.none.fl_str_mv COMPUTERS & INDUSTRIAL ENGINEERING, 172 (Parte A), 108611-1.
https://doi.org/10.1016/j.cie.2022.108611
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
application/pdf
dc.publisher.none.fl_str_mv ScienceDirect
publisher.none.fl_str_mv ScienceDirect
dc.source.none.fl_str_mv reponame:idUS. Depósito de Investigación de la Universidad de Sevilla
instname:Universidad de Sevilla (US)
instname_str Universidad de Sevilla (US)
reponame_str idUS. Depósito de Investigación de la Universidad de Sevilla
collection idUS. Depósito de Investigación de la Universidad de Sevilla
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869411491982082048
score 15,301629

Registos relacionados