Effect of label noise in the complexity of classification problems

Página do item simplificado
dc.contributor.authorGarcia, Luis P. F.
dc.contributor.authorCarvalho, Andre C. P. L. F. de
dc.contributor.authorLorena, Ana C. [UNIFESP]
dc.contributor.institutionUniversidade de São Paulo (USP)
dc.contributor.institutionUniversidade Federal de São Paulo (UNIFESP)
dc.date.accessioned2016-01-24T14:40:35Z
dc.date.available2016-01-24T14:40:35Z
dc.date.issued2015-07-21
dc.description.abstractNoisy data are common in real-World problems and may have several causes, like inaccuracies, distortions or contamination during data collection, storage and/or transmission. the presence of noise in data can affect the complexity of classification problems, making the discrimination of objects from different classes more difficult, and requiring more complex decision boundaries for data separation. in this paper, we investigate how noise affects the complexity of classification problems, by monitoring the sensitivity of several indices of data complexity in the presence of different label noise levels. To characterize the complexity of a classification dataset, we use geometric, statistical and structural measures extracted from data. the experimental results show that some measures are more sensitive than others to the addition of noise in a dataset These measures can be used in the development of new preprocessing techniques for noise identification and novel label noise tolerant algorithms. We thereby show preliminary results on a new filter for noise identification, which is based on two of the complexity measures which were more sensitive to the presence of label noise. (C) 2015 Elsevier B.V. All rights reserved.en
dc.description.affiliationUniv São Paulo, Inst Ciencias Matemat & Comp, BR-13560970 São Paulo, Brazil
dc.description.affiliationUniversidade Federal de São Paulo, Inst Ciencia & Tecnol, BR-12231280 São Paulo, Brazil
dc.description.affiliationUnifespUniversidade Federal de São Paulo, Inst Ciencia & Tecnol, BR-12231280 São Paulo, Brazil
dc.description.sourceWeb of Science
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.description.sponsorshipCoordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
dc.format.extent108-119
dc.identifierhttp://dx.doi.org/10.1016/j.neucom.2014.10.085
dc.identifier.citationNeurocomputing. Amsterdam: Elsevier B.V., v. 160, p. 108-119, 2015.
dc.identifier.doi10.1016/j.neucom.2014.10.085
dc.identifier.issn0925-2312
dc.identifier.urihttp://repositorio.unifesp.br/handle/11600/39154
dc.identifier.wosWOS:000354139100010
dc.language.isoeng
dc.publisherElsevier B.V.
dc.relation.ispartofNeurocomputing
dc.rightsinfo:eu-repo/semantics/restrictedAccess
dc.rights.licensehttp://www.elsevier.com/about/open-access/open-access-policies/article-posting-policy
dc.subjectClassificationen
dc.subjectLabel noiseen
dc.subjectComplexity measuresen
dc.subjectNoise Filteren
dc.titleEffect of label noise in the complexity of classification problemsen
dc.typeinfo:eu-repo/semantics/article
Arquivos
Coleções
ICT - Artigos