High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems
Ano de defesa: | 2023 |
---|---|
Autor(a) principal: | |
Orientador(a): | |
Banca de defesa: | , , , |
Tipo de documento: | Tese |
Tipo de acesso: | Acesso aberto |
Idioma: | eng |
Instituição de defesa: |
Universidade Federal de Minas Gerais
|
Programa de Pós-Graduação: |
Programa de Pós-Graduação em Engenharia Elétrica
|
Departamento: |
ENG - DEPARTAMENTO DE ENGENHARIA ELÉTRICA
|
País: |
Brasil
|
Palavras-chave em Português: | |
Link de acesso: | http://hdl.handle.net/1843/58467 |
Resumo: | Fuzzy Cognitive Maps (FCMs) have emerged as interpretable Fuzzy Time Series (FTS) methods used in a variety of forecasting applications due to their interesting features. Constructing the structure of FCMs and extracting weighted connections among the concepts compose the crux contribution of the proposed FCM-based approaches in the literature. Despite the success of the proposed methodologies, there are still some gaps and limitations in this domain. To cover some of these challenges, this thesis introduces new forecasting techniques based on FCMs to predict univariate and multivariate time series focusing on both aspects including designing the new architecture and speeding up the training phase. Thus, the main contribution of this thesis is to introduce novel forecasting techniques by merging FTS and FCMs to generate randomized high-order FCM (R-HFCM) as reservoir computing models for the first time in the literature. R-HFCM is a kind of ESN where the reservoir layer consists of a group of sub-reservoirs such that the weights inside each sub-reservoir are randomly chosen according to the ESN weight initialization. The computational experiments demonstrate that R-HFCM outperforms in terms of both accuracy and training speed when compared to the traditional FCMs trained via evolutionary algorithms like genetic algorithm (GA). To fill the absence of Multiple-Input Multiple-Output(MIMO) models, extensions of the univariate R-HFCM method are presented to handle low-dimensional and high-dimensional time series forecasting. It is worth noting that in both MIMO methods, only the output layer is trainable using the time-effective least squares method. The proposed methods obtained promising and competitive results compared with a range variety of deep learning and machine learning methods in terms of accuracy and parsimony. |
id |
UFMG_394ee2fbe16ccd079d73f57254eab16e |
---|---|
oai_identifier_str |
oai:repositorio.ufmg.br:1843/58467 |
network_acronym_str |
UFMG |
network_name_str |
Repositório Institucional da UFMG |
repository_id_str |
|
spelling |
Frederico Gadelha Guimarãeshttp://lattes.cnpq.br/2472681535872194Petrônio Cândido de Lima e SilvaGuilherme de Alencar BarretoTatiane Nogueira RiosHugo Valadares SiqueiraRodrigo César Pedrosa Silvahttp://lattes.cnpq.br/5764595026529407Omid Orang2023-09-05T18:32:30Z2023-09-05T18:32:30Z2023-06-30http://hdl.handle.net/1843/584670000-0002-4077-3775Fuzzy Cognitive Maps (FCMs) have emerged as interpretable Fuzzy Time Series (FTS) methods used in a variety of forecasting applications due to their interesting features. Constructing the structure of FCMs and extracting weighted connections among the concepts compose the crux contribution of the proposed FCM-based approaches in the literature. Despite the success of the proposed methodologies, there are still some gaps and limitations in this domain. To cover some of these challenges, this thesis introduces new forecasting techniques based on FCMs to predict univariate and multivariate time series focusing on both aspects including designing the new architecture and speeding up the training phase. Thus, the main contribution of this thesis is to introduce novel forecasting techniques by merging FTS and FCMs to generate randomized high-order FCM (R-HFCM) as reservoir computing models for the first time in the literature. R-HFCM is a kind of ESN where the reservoir layer consists of a group of sub-reservoirs such that the weights inside each sub-reservoir are randomly chosen according to the ESN weight initialization. The computational experiments demonstrate that R-HFCM outperforms in terms of both accuracy and training speed when compared to the traditional FCMs trained via evolutionary algorithms like genetic algorithm (GA). To fill the absence of Multiple-Input Multiple-Output(MIMO) models, extensions of the univariate R-HFCM method are presented to handle low-dimensional and high-dimensional time series forecasting. It is worth noting that in both MIMO methods, only the output layer is trainable using the time-effective least squares method. The proposed methods obtained promising and competitive results compared with a range variety of deep learning and machine learning methods in terms of accuracy and parsimony.Os mapas cognitivos nebulosos (FCM, do inglês Fuzzy Cognitive Maps) surgiram como métodos interpretáveis das Séries Temporais Nebulosas (FTS, do inglês Fuzzy Time Series) para uma variedade de aplicações no campo de previsão. A construção da estrutura dos FCMs e a extração das conexões ponderadas entre os conceitos compõem a contribuição central das abordagens baseadas em FCMs na literatura. Apesar do sucesso das metodologias propostas, ainda existem algumas lacunas e limitações nesse domínio. Para cobrir alguns desses desafios, esta tese apresenta novas técnicas de previsão baseadas em FCMs para prever séries temporais univariadas e multivariadas, focando no design da nova arquitetura e na aceleração da fase de treinamento. Assim, a principal contribuição desta tese é introduzir novas técnicas de previsão pela fusão de FTS e FCMs para gerar FCMs aleatórios de alta ordem (R-HFCM, do inglês Randomized High-Order FCM) como modelos de computação de reservatório pela primeira vez na literatura. O R-HFCM é um tipo de rede de estado de eco (ESN, do inglês Echo State Network ), onde a camada do reservatório consiste em um grupo de sub-reservatórios de tal forma que os pesos dentro de cada sub-reservatório são escolhidos aleatoriamente de acordo com a inicialização de pesos do ESN. Os experimentos computacionais demonstram que o R-HFCM supera em termos de precisão e velocidade de treinamento quando comparado aos FCMs tradicionais treinados por algoritmos evolutivos como o algoritmo genético (GA, do inglês Genethic Algorithm). Para preencher a ausência de modelos de Entrada Múltipla e Saída Múltipla (MIMO, do inglês Multiple-Input Multiple-Output), extensões do método R-HFCM univariado foram apresentadas para lidar com a previsão de séries temporais de baixa e alta dimensionalidade. Vale ressaltar que, em ambos os métodos MIMO, apenas a camada de saída é treinável utilizando o método dos mínimos quadrados por ser de baixo custo computacional. Os métodos propostos obtiveram resultados promissores e competitivos em comparação com uma variedade de métodos de aprendizado profundo e aprendizado de máquina em termos de precisão e parcimônia.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/3.0/pt/info:eu-repo/semantics/openAccessEngenharia elétricaSéries temporaisMapas cognitivos (Psicologia)AlgoritmosAlgoritmos genéticosRedes neurais (Computação)Aprendizado do computadorMínimos quadradosAnálise de componentes principaisSistemas de comunicação sem fioKernel, Funções deTime series forecastingRandomized fuzzy cognitive mapsReservoir computingEcho state networkMultiple-input multiple-outputHigh-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systemsinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesisreponame:Repositório Institucional da UFMGinstname:Universidade Federal de Minas Gerais (UFMG)instacron:UFMGORIGINALPhD_thesis_Omid_final_template_.pdfPhD_thesis_Omid_final_template_.pdfDoctorate Thesis Omid Orangapplication/pdf3923257https://repositorio.ufmg.br/bitstream/1843/58467/1/PhD_thesis_Omid_final_template_.pdfddb8159ab9b685f1fd63fcbf47f813adMD51CC-LICENSElicense_rdflicense_rdfapplication/rdf+xml; charset=utf-8914https://repositorio.ufmg.br/bitstream/1843/58467/2/license_rdff9944a358a0c32770bd9bed185bb5395MD52LICENSElicense.txtlicense.txttext/plain; charset=utf-82118https://repositorio.ufmg.br/bitstream/1843/58467/3/license.txtcda590c95a0b51b4d15f60c9642ca272MD531843/584672023-09-05 15:32:30.964oai:repositorio.ufmg.br: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ório de PublicaçõesPUBhttps://repositorio.ufmg.br/oaiopendoar:2023-09-05T18:32:30Repositório Institucional da UFMG - Universidade Federal de Minas Gerais (UFMG)false |
dc.title.pt_BR.fl_str_mv |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
title |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
spellingShingle |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems Omid Orang Time series forecasting Randomized fuzzy cognitive maps Reservoir computing Echo state network Multiple-input multiple-output Engenharia elétrica Séries temporais Mapas cognitivos (Psicologia) Algoritmos Algoritmos genéticos Redes neurais (Computação) Aprendizado do computador Mínimos quadrados Análise de componentes principais Sistemas de comunicação sem fio Kernel, Funções de |
title_short |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
title_full |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
title_fullStr |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
title_full_unstemmed |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
title_sort |
High-order fuzzy cognitive maps and randomized networks for time series and nonlinear dynamical systems |
author |
Omid Orang |
author_facet |
Omid Orang |
author_role |
author |
dc.contributor.advisor1.fl_str_mv |
Frederico Gadelha Guimarães |
dc.contributor.advisor1Lattes.fl_str_mv |
http://lattes.cnpq.br/2472681535872194 |
dc.contributor.advisor-co1.fl_str_mv |
Petrônio Cândido de Lima e Silva |
dc.contributor.referee1.fl_str_mv |
Guilherme de Alencar Barreto |
dc.contributor.referee2.fl_str_mv |
Tatiane Nogueira Rios |
dc.contributor.referee3.fl_str_mv |
Hugo Valadares Siqueira |
dc.contributor.referee4.fl_str_mv |
Rodrigo César Pedrosa Silva |
dc.contributor.authorLattes.fl_str_mv |
http://lattes.cnpq.br/5764595026529407 |
dc.contributor.author.fl_str_mv |
Omid Orang |
contributor_str_mv |
Frederico Gadelha Guimarães Petrônio Cândido de Lima e Silva Guilherme de Alencar Barreto Tatiane Nogueira Rios Hugo Valadares Siqueira Rodrigo César Pedrosa Silva |
dc.subject.por.fl_str_mv |
Time series forecasting Randomized fuzzy cognitive maps Reservoir computing Echo state network Multiple-input multiple-output |
topic |
Time series forecasting Randomized fuzzy cognitive maps Reservoir computing Echo state network Multiple-input multiple-output Engenharia elétrica Séries temporais Mapas cognitivos (Psicologia) Algoritmos Algoritmos genéticos Redes neurais (Computação) Aprendizado do computador Mínimos quadrados Análise de componentes principais Sistemas de comunicação sem fio Kernel, Funções de |
dc.subject.other.pt_BR.fl_str_mv |
Engenharia elétrica Séries temporais Mapas cognitivos (Psicologia) Algoritmos Algoritmos genéticos Redes neurais (Computação) Aprendizado do computador Mínimos quadrados Análise de componentes principais Sistemas de comunicação sem fio Kernel, Funções de |
description |
Fuzzy Cognitive Maps (FCMs) have emerged as interpretable Fuzzy Time Series (FTS) methods used in a variety of forecasting applications due to their interesting features. Constructing the structure of FCMs and extracting weighted connections among the concepts compose the crux contribution of the proposed FCM-based approaches in the literature. Despite the success of the proposed methodologies, there are still some gaps and limitations in this domain. To cover some of these challenges, this thesis introduces new forecasting techniques based on FCMs to predict univariate and multivariate time series focusing on both aspects including designing the new architecture and speeding up the training phase. Thus, the main contribution of this thesis is to introduce novel forecasting techniques by merging FTS and FCMs to generate randomized high-order FCM (R-HFCM) as reservoir computing models for the first time in the literature. R-HFCM is a kind of ESN where the reservoir layer consists of a group of sub-reservoirs such that the weights inside each sub-reservoir are randomly chosen according to the ESN weight initialization. The computational experiments demonstrate that R-HFCM outperforms in terms of both accuracy and training speed when compared to the traditional FCMs trained via evolutionary algorithms like genetic algorithm (GA). To fill the absence of Multiple-Input Multiple-Output(MIMO) models, extensions of the univariate R-HFCM method are presented to handle low-dimensional and high-dimensional time series forecasting. It is worth noting that in both MIMO methods, only the output layer is trainable using the time-effective least squares method. The proposed methods obtained promising and competitive results compared with a range variety of deep learning and machine learning methods in terms of accuracy and parsimony. |
publishDate |
2023 |
dc.date.accessioned.fl_str_mv |
2023-09-05T18:32:30Z |
dc.date.available.fl_str_mv |
2023-09-05T18:32:30Z |
dc.date.issued.fl_str_mv |
2023-06-30 |
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/1843/58467 |
dc.identifier.orcid.pt_BR.fl_str_mv |
0000-0002-4077-3775 |
url |
http://hdl.handle.net/1843/58467 |
identifier_str_mv |
0000-0002-4077-3775 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
http://creativecommons.org/licenses/by/3.0/pt/ info:eu-repo/semantics/openAccess |
rights_invalid_str_mv |
http://creativecommons.org/licenses/by/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/1843/58467/1/PhD_thesis_Omid_final_template_.pdf https://repositorio.ufmg.br/bitstream/1843/58467/2/license_rdf https://repositorio.ufmg.br/bitstream/1843/58467/3/license.txt |
bitstream.checksum.fl_str_mv |
ddb8159ab9b685f1fd63fcbf47f813ad f9944a358a0c32770bd9bed185bb5395 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_ |
1793891059950944256 |