Data complexity meta-features for regression problems

Página do item simplificado
dc.citation.issue1
dc.citation.volume107
dc.contributor.authorLorena, Ana C. [UNIFESP]
dc.contributor.authorMaciel, Aron I. [UNIFESP]
dc.contributor.authorde Miranda, Pericles B. C.
dc.contributor.authorCosta, Ivan G.
dc.contributor.authorPrudencio, Ricardo B. C.
dc.coverageDordrecht
dc.date.accessioned2020-07-02T18:52:11Z
dc.date.available2020-07-02T18:52:11Z
dc.date.issued2018
dc.description.abstractIn meta-learning, classification problems can be described by a variety of features, including complexity measures. These measures allow capturing the complexity of the frontier that separates the classes. For regression problems, on the other hand, there is a lack of such type of measures. This paper presents and analyses measures devoted to estimate the complexity of the function that should fitted to the data in regression problems. As case studies, they are employed as meta-features in three meta-learning setups: (i) the first one predicts the regression function type of some synthetic datasets; (ii) the second one is designed to tune the parameter values of support vector regressors; and (iii) the third one aims to predict the performance of various regressors for a given dataset. The results show the suitability of the new measures to describe the regression datasets and their utility in the meta-learning tasks considered. In cases (ii) and (iii) the achieved results are also similar or better than those obtained by the use of classical meta-features in meta-learning.en
dc.description.affiliationUniv Fed Sao Paulo, Inst Ciencia Tecnol, Unidade Parque Tecnol, BR-12247014 Sao Jose Dos Campos, SP, Brazil
dc.description.affiliationUniv Fed Pernambuco, Ctr Informat, BR-50740560 Recife, PE, Brazil
dc.description.affiliationRhein Westfal TH Aachen, IZKF Res Grp Bioinformat, Aachen, Germany
dc.description.affiliationUnifespUniv Fed Sao Paulo, Inst Ciencia Tecnol, Unidade Parque Tecnol, BR-12247014 Sao Jose Dos Campos, SP, Brazil
dc.description.sourceWeb of Science
dc.description.sponsorshipFAPESP
dc.description.sponsorshipCNPq
dc.description.sponsorshipCAPES
dc.description.sponsorshipDAAD
dc.description.sponsorshipIZKF Aachen
dc.description.sponsorshipIDFAPESP: 2012/22608-8
dc.description.sponsorshipIDCNPq: 482222/2013-1
dc.description.sponsorshipIDCNPq: 308858/2014-0
dc.description.sponsorshipIDCNPq: 305611/2015-1
dc.format.extent209-246
dc.identifierhttp://dx.doi.org/10.1007/s10994-017-5681-1
dc.identifier.citationMachine Learning. Dordrecht, v. 107, n. 1, p. 209-246, 2018.
dc.identifier.doi10.1007/s10994-017-5681-1
dc.identifier.fileWOS000419684700008.pdf
dc.identifier.issn0885-6125
dc.identifier.urihttps://repositorio.unifesp.br/handle/11600/53930
dc.identifier.wosWOS:000419684700008
dc.language.isoeng
dc.publisherSpringer
dc.relation.ispartofMachine Learning
dc.rightsinfo:eu-repo/semantics/restrictedAccess
dc.subjectMeta-learningen
dc.subjectMeta-featuresen
dc.subjectComplexity measuresen
dc.titleData complexity meta-features for regression problemsen
dc.typeinfo:eu-repo/semantics/article
Arquivos
Coleções
ICT - Artigos