Logotipo do repositório
 

Publicação:
HOW FAR do WE GET USING MACHINE LEARNING BLACK-BOXES?

dc.contributor.authorRocha, Anderson
dc.contributor.authorPapa, João Paulo [UNESP]
dc.contributor.authorMeira, Luis A. A.
dc.contributor.institutionUniversidade Estadual de Campinas (UNICAMP)
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.date.accessioned2014-05-20T13:25:58Z
dc.date.available2014-05-20T13:25:58Z
dc.date.issued2012-03-01
dc.description.abstractWith several good research groups actively working in machine learning (ML) approaches, we have now the concept of self-containing machine learning solutions that oftentimes work out-of-the-box leading to the concept of ML black-boxes. Although it is important to have such black-boxes helping researchers to deal with several problems nowadays, it comes with an inherent problem increasingly more evident: we have observed that researchers and students are progressively relying on ML black-boxes and, usually, achieving results without knowing the machinery of the classifiers. In this regard, this paper discusses the use of machine learning black-boxes and poses the question of how far we can get using these out-of-the-box solutions instead of going deeper into the machinery of the classifiers. The paper focuses on three aspects of classifiers: (1) the way they compare examples in the feature space; (2) the impact of using features with variable dimensionality; and (3) the impact of using binary classifiers to solve a multi-class problem. We show how knowledge about the classifier's machinery can improve the results way beyond out-of-the-box machine learning solutions.en
dc.description.affiliationUniv Campinas UNICAMP, Inst Comp, BR-13083852 Campinas, SP, Brazil
dc.description.affiliationUNESP Univ Estadual Paulista, Dept Comp Sci, BR-17033360 Bauru, SP, Brazil
dc.description.affiliationUniv Campinas UNICAMP, Fac Technol, BR-13484332 Limeira, SP, Brazil
dc.description.affiliationUnespUNESP Univ Estadual Paulista, Dept Comp Sci, BR-17033360 Bauru, SP, Brazil
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipUniversity of Campinas PAPDIC Grant
dc.description.sponsorshipConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.description.sponsorshipMicrosoft Research
dc.description.sponsorshipIdFAPESP: 09/16206-1
dc.description.sponsorshipIdFAPESP: 10/05647-4
dc.description.sponsorshipIdUniversity of Campinas PAPDIC Grant: 519.292-340/10
dc.format.extent23
dc.identifierhttp://dx.doi.org/10.1142/S0218001412610010
dc.identifier.citationInternational Journal of Pattern Recognition and Artificial Intelligence. Singapore: World Scientific Publ Co Pte Ltd, v. 26, n. 2, p. 23, 2012.
dc.identifier.doi10.1142/S0218001412610010
dc.identifier.issn0218-0014
dc.identifier.lattes9039182932747194
dc.identifier.urihttp://hdl.handle.net/11449/8295
dc.identifier.wosWOS:000308104300007
dc.language.isoeng
dc.publisherWorld Scientific Publ Co Pte Ltd
dc.relation.ispartofInternational Journal of Pattern Recognition and Artificial Intelligence
dc.relation.ispartofjcr1.029
dc.relation.ispartofsjr0,315
dc.rights.accessRightsAcesso restrito
dc.sourceWeb of Science
dc.subjectMachine learning black-boxesen
dc.subjectbinary to multi-class classifiersen
dc.subjectsupport vector machinesen
dc.subjectoptimum-path foresten
dc.subjectvisual wordsen
dc.subjectK-nearest neighborsen
dc.titleHOW FAR do WE GET USING MACHINE LEARNING BLACK-BOXES?en
dc.typeArtigo
dcterms.licensehttp://www.worldscientific.com/page/authors/author-rights
dcterms.rightsHolderWorld Scientific Publ Co Pte Ltd
dspace.entity.typePublication
unesp.author.lattes9039182932747194
unesp.author.orcid0000-0002-6494-7514[2]
unesp.campusUniversidade Estadual Paulista (UNESP), Faculdade de Ciências, Baurupt
unesp.departmentComputação - FCpt

Arquivos

Licença do Pacote

Agora exibindo 1 - 2 de 2
Carregando...
Imagem de Miniatura
Nome:
license.txt
Tamanho:
1.71 KB
Formato:
Item-specific license agreed upon to submission
Descrição:
Carregando...
Imagem de Miniatura
Nome:
license.txt
Tamanho:
1.71 KB
Formato:
Item-specific license agreed upon to submission
Descrição: