Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph
Autor(a) principal: | |
---|---|
Data de Publicação: | 2022 |
Tipo de documento: | Tese |
Idioma: | eng |
Título da fonte: | Repositório Institucional da UFMG |
Texto Completo: | http://hdl.handle.net/1843/47406 |
Resumo: | Online Learning of non-stationary data streams is characterized by changes in the data generating function, which may impact the predictive performance of a model. Therefore, classifiers capable of adapting to such situations constitute a viable solution. Generally, such models rely on hyperparameters that need to be previously configured. A different task that also presents issues, concerning setting of hyperparameters, is the learning with one-class classifiers, in which the information from only one class is used to establish a decision boundary. The main proposal of this thesis is to use the structural information from a data set to define classifiers, in the two Learning Paradigms previously discussed. This goal is achieved by exploring the fact that an Independent Dominating Set, when induced from the Gabriel graph, tends by definition to result in a subset of dominating points, with representative characteristic of the original set. Thus, an Independent Dominating Set algorithm that neither requires setting hyperparameters nor the use of any optimization method to find a solution is proposed, as well as an online updating procedure for the Gabriel graph. These two methods are used to define the hypeparameters of models based on Radial functions: a KDE estimator for the online scenario and an RBF network as a one-class classifier. This graph dominance approach results in an appropriate and distributed number of Radial functions, in the input domain, and a stable radius that cover the training points and leads to a classifier with appropriate Capacity. An algorithm based on the independent dominating set of the Gabriel graph is also proposed to extract representative subsets from large data sets. This thesis also presents an online training method for a regularized SLFN network that continually maintains the learning process. The method uses an adaptive window to mitigate the impact of concept drifts. These methods were tested with synthetic data sets and with data from a real industrial process. |
id |
UFMG_ff972b77497b27952e7f72cb0abaafc3 |
---|---|
oai_identifier_str |
oai:repositorio.ufmg.br:1843/47406 |
network_acronym_str |
UFMG |
network_name_str |
Repositório Institucional da UFMG |
repository_id_str |
|
spelling |
Antônio de Pádua Bragahttp://lattes.cnpq.br/1130012055294645Walmir Matos CaminhasJanier Arias GarcíaAluízio Fausto Ribeiro AraújoRaul Fonseca Netohttp://lattes.cnpq.br/9141944384996872Wagner José de Alvarenga Júnior2022-11-23T17:44:47Z2022-11-23T17:44:47Z2022-07-14http://hdl.handle.net/1843/474060000-0002-4870-9524Online Learning of non-stationary data streams is characterized by changes in the data generating function, which may impact the predictive performance of a model. Therefore, classifiers capable of adapting to such situations constitute a viable solution. Generally, such models rely on hyperparameters that need to be previously configured. A different task that also presents issues, concerning setting of hyperparameters, is the learning with one-class classifiers, in which the information from only one class is used to establish a decision boundary. The main proposal of this thesis is to use the structural information from a data set to define classifiers, in the two Learning Paradigms previously discussed. This goal is achieved by exploring the fact that an Independent Dominating Set, when induced from the Gabriel graph, tends by definition to result in a subset of dominating points, with representative characteristic of the original set. Thus, an Independent Dominating Set algorithm that neither requires setting hyperparameters nor the use of any optimization method to find a solution is proposed, as well as an online updating procedure for the Gabriel graph. These two methods are used to define the hypeparameters of models based on Radial functions: a KDE estimator for the online scenario and an RBF network as a one-class classifier. This graph dominance approach results in an appropriate and distributed number of Radial functions, in the input domain, and a stable radius that cover the training points and leads to a classifier with appropriate Capacity. An algorithm based on the independent dominating set of the Gabriel graph is also proposed to extract representative subsets from large data sets. This thesis also presents an online training method for a regularized SLFN network that continually maintains the learning process. The method uses an adaptive window to mitigate the impact of concept drifts. These methods were tested with synthetic data sets and with data from a real industrial process.O Aprendizado Online de dados não estacionários é caracterizado por mudanças na função geradora dos dados, com possível impacto sobre o desempenho de um modelo preditor. Por isto, classificadores que apresentam Capacidade apropriada ao longo das predições são uma possível solução. Porém, tais modelos geralmente possuem hiperparâmetros que necessitam ser definidos previamente. Uma tarefa diferente desta, que também possui desafios relacionados a determinação da Capacidade de um modelo, é o aprendizado realizado com classificadores de classe única. Neste modelo, a superfície de decisão é induzida a partir dos dados referentes a uma única classe. A proposta principal desta tese está centrada no uso da informação da estrutura dos dados para se definir classificadores nos dois Paradigmas de Aprendizados anteriores. Para isto, utiliza-se de um Conjunto Dominante Independente do grafo de Gabriel, o qual tende por definição a ser um subconjunto de pontos dominantes distribuídos e com característica de representatividade do conjunto original. Desta forma, é proposto um algoritmo para a obtenção de um conjunto dominante independente, o qual não necessita de configurar parâmetros e também não utiliza métodos de otimização para achar a solução. É também proposto, uma abordagem para se atualizar, de forma online, o grafo de Gabriel com um novo ponto. Estes dois métodos são usados na definição hiperparamétrica de modelos com funções Radiais: um estimador KDE empregado no cenário de aprendizado online e uma rede RBF utilizada como classificado de classe única. Além destes, é proposto um algoritmo baseado no conjunto dominante independente e no grafo de Gabriel, cuja finalidade é extrair subconjuntos representativos de um conjunto original que possua muito pontos. Esta tese apresenta ainda um método de treinamento online para uma rede SLFN. O processo contínuo de treinamento utiliza uma janela adaptativa para atenuar o impacto causado por mudanças de conceito. Esses métodos foram testados com conjuntos de dados sintéticos e com dados de um processo industrial real.CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível SuperiorengUniversidade Federal de Minas GeraisPrograma de Pós-Graduação em Engenharia ElétricaUFMGBrasilENG - DEPARTAMENTO DE ENGENHARIA ELÉTRICAhttp://creativecommons.org/licenses/by-nc/3.0/pt/info:eu-repo/semantics/openAccessEngenharia elétricaAprendizado do computadorTeoria dos grafosOnline learnigGabriel graphDominating setKDERBFOne-class classifierSlFNDominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graphConjuntos dominantes do grafo de Gabriel : uma abordagem para classificadores de classe única e para o aprendizado onlineinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesisreponame:Repositório Institucional da UFMGinstname:Universidade Federal de Minas Gerais (UFMG)instacron:UFMGCC-LICENSElicense_rdflicense_rdfapplication/rdf+xml; charset=utf-8920https://repositorio.ufmg.br/bitstream/1843/47406/2/license_rdf33b8016dc5c4681c1e7a582a4161162cMD52ORIGINALTexto_Repositorio.pdfTexto_Repositorio.pdfTexto da teseapplication/pdf62254952https://repositorio.ufmg.br/bitstream/1843/47406/4/Texto_Repositorio.pdf7496f565cb0f57340d0ca454205bf483MD54LICENSElicense.txtlicense.txttext/plain; charset=utf-82118https://repositorio.ufmg.br/bitstream/1843/47406/5/license.txtcda590c95a0b51b4d15f60c9642ca272MD551843/474062022-11-23 14:44:48.112oai:repositorio.ufmg.br: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ório de PublicaçõesPUBhttps://repositorio.ufmg.br/oaiopendoar:2022-11-23T17:44:48Repositório Institucional da UFMG - Universidade Federal de Minas Gerais (UFMG)false |
dc.title.pt_BR.fl_str_mv |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
dc.title.alternative.pt_BR.fl_str_mv |
Conjuntos dominantes do grafo de Gabriel : uma abordagem para classificadores de classe única e para o aprendizado online |
title |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
spellingShingle |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph Wagner José de Alvarenga Júnior Online learnig Gabriel graph Dominating set KDE RBF One-class classifier SlFN Engenharia elétrica Aprendizado do computador Teoria dos grafos |
title_short |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
title_full |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
title_fullStr |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
title_full_unstemmed |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
title_sort |
Dominating sets of the Gabriel Graph : na approach for one-class and online learning classifiers Gabriel graph |
author |
Wagner José de Alvarenga Júnior |
author_facet |
Wagner José de Alvarenga Júnior |
author_role |
author |
dc.contributor.advisor1.fl_str_mv |
Antônio de Pádua Braga |
dc.contributor.advisor1Lattes.fl_str_mv |
http://lattes.cnpq.br/1130012055294645 |
dc.contributor.referee1.fl_str_mv |
Walmir Matos Caminhas |
dc.contributor.referee2.fl_str_mv |
Janier Arias García |
dc.contributor.referee3.fl_str_mv |
Aluízio Fausto Ribeiro Araújo |
dc.contributor.referee4.fl_str_mv |
Raul Fonseca Neto |
dc.contributor.authorLattes.fl_str_mv |
http://lattes.cnpq.br/9141944384996872 |
dc.contributor.author.fl_str_mv |
Wagner José de Alvarenga Júnior |
contributor_str_mv |
Antônio de Pádua Braga Walmir Matos Caminhas Janier Arias García Aluízio Fausto Ribeiro Araújo Raul Fonseca Neto |
dc.subject.por.fl_str_mv |
Online learnig Gabriel graph Dominating set KDE RBF One-class classifier SlFN |
topic |
Online learnig Gabriel graph Dominating set KDE RBF One-class classifier SlFN Engenharia elétrica Aprendizado do computador Teoria dos grafos |
dc.subject.other.pt_BR.fl_str_mv |
Engenharia elétrica Aprendizado do computador Teoria dos grafos |
description |
Online Learning of non-stationary data streams is characterized by changes in the data generating function, which may impact the predictive performance of a model. Therefore, classifiers capable of adapting to such situations constitute a viable solution. Generally, such models rely on hyperparameters that need to be previously configured. A different task that also presents issues, concerning setting of hyperparameters, is the learning with one-class classifiers, in which the information from only one class is used to establish a decision boundary. The main proposal of this thesis is to use the structural information from a data set to define classifiers, in the two Learning Paradigms previously discussed. This goal is achieved by exploring the fact that an Independent Dominating Set, when induced from the Gabriel graph, tends by definition to result in a subset of dominating points, with representative characteristic of the original set. Thus, an Independent Dominating Set algorithm that neither requires setting hyperparameters nor the use of any optimization method to find a solution is proposed, as well as an online updating procedure for the Gabriel graph. These two methods are used to define the hypeparameters of models based on Radial functions: a KDE estimator for the online scenario and an RBF network as a one-class classifier. This graph dominance approach results in an appropriate and distributed number of Radial functions, in the input domain, and a stable radius that cover the training points and leads to a classifier with appropriate Capacity. An algorithm based on the independent dominating set of the Gabriel graph is also proposed to extract representative subsets from large data sets. This thesis also presents an online training method for a regularized SLFN network that continually maintains the learning process. The method uses an adaptive window to mitigate the impact of concept drifts. These methods were tested with synthetic data sets and with data from a real industrial process. |
publishDate |
2022 |
dc.date.accessioned.fl_str_mv |
2022-11-23T17:44:47Z |
dc.date.available.fl_str_mv |
2022-11-23T17:44:47Z |
dc.date.issued.fl_str_mv |
2022-07-14 |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/doctoralThesis |
format |
doctoralThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/1843/47406 |
dc.identifier.orcid.pt_BR.fl_str_mv |
0000-0002-4870-9524 |
url |
http://hdl.handle.net/1843/47406 |
identifier_str_mv |
0000-0002-4870-9524 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
http://creativecommons.org/licenses/by-nc/3.0/pt/ info:eu-repo/semantics/openAccess |
rights_invalid_str_mv |
http://creativecommons.org/licenses/by-nc/3.0/pt/ |
eu_rights_str_mv |
openAccess |
dc.publisher.none.fl_str_mv |
Universidade Federal de Minas Gerais |
dc.publisher.program.fl_str_mv |
Programa de Pós-Graduação em Engenharia Elétrica |
dc.publisher.initials.fl_str_mv |
UFMG |
dc.publisher.country.fl_str_mv |
Brasil |
dc.publisher.department.fl_str_mv |
ENG - DEPARTAMENTO DE ENGENHARIA ELÉTRICA |
publisher.none.fl_str_mv |
Universidade Federal de Minas Gerais |
dc.source.none.fl_str_mv |
reponame:Repositório Institucional da UFMG instname:Universidade Federal de Minas Gerais (UFMG) instacron:UFMG |
instname_str |
Universidade Federal de Minas Gerais (UFMG) |
instacron_str |
UFMG |
institution |
UFMG |
reponame_str |
Repositório Institucional da UFMG |
collection |
Repositório Institucional da UFMG |
bitstream.url.fl_str_mv |
https://repositorio.ufmg.br/bitstream/1843/47406/2/license_rdf https://repositorio.ufmg.br/bitstream/1843/47406/4/Texto_Repositorio.pdf https://repositorio.ufmg.br/bitstream/1843/47406/5/license.txt |
bitstream.checksum.fl_str_mv |
33b8016dc5c4681c1e7a582a4161162c 7496f565cb0f57340d0ca454205bf483 cda590c95a0b51b4d15f60c9642ca272 |
bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 MD5 |
repository.name.fl_str_mv |
Repositório Institucional da UFMG - Universidade Federal de Minas Gerais (UFMG) |
repository.mail.fl_str_mv |
|
_version_ |
1803589213258514432 |