Advances in time series analysis: segmentation, prediction, and unsupervised tuning of drift detectors

Página do item simplificado
dc.contributor.advisorMelo, Leonimer Flávio de
dc.contributor.authorSilva, Ricardo Petri
dc.contributor.bancaFelinto, Alan Salvany
dc.contributor.bancaKaster, Daniel dos Santos
dc.contributor.bancaNaozuka, Gustavo Taiji
dc.contributor.bancaGuido, Rodrigo Capobianco
dc.coverage.extent88 p.
dc.coverage.spatialLondrina
dc.date.accessioned2026-09-24T12:11:38Z
dc.date.available2026-09-24T12:11:38Z
dc.date.issued2024-07-12
dc.description.abstractTime series analysis proves highly useful and widely applicable, enabling reliable forecasting and informed decision-making across various fields. This thesis describes the fundamental importance of time series analysis and introduces advanced techniques aimed at enhancing the structure of time series data to improve predictive accuracy. Given the challenges associated with concept drifts—where the statistical properties of a signal change unpredictably—this study also emphasizes the need for innovative, unsupervised approaches to optimize drift detectors. The thesis presents two novel segmentation methods designed to refine data processing based on stationarity analysis, thus facilitating more accurate forecasts. Additionally, it develops a new method to enhance drift detectors using unsupervised approach, reducing the need for manual tuning and deep prior knowledge of the data. These contributions significantly advance the field of time series analysis, providing robust, adaptable, and efficient solutions for applications where precise predictions are vital. Our results demonstrate that the segmentation process enhances the prediction process for both Long Short-term Memory and Temporal Convolutional Networks. Nonetheless, our unsupervised tuning outperforms the default configuration in drift detectors in most evaluated cases.
dc.description.abstractother1Resumo
dc.identifier.urihttps://repositorio.uel.br/handle/123456789/20013
dc.language.isoeng
dc.relation.departamentCTU - Departamento de Engenharia Elétrica
dc.relation.institutionnameUniversidade Estadual de Londrina - UEL
dc.relation.ppgnamePrograma de Pós-Graduação em Engenharia Elétrica
dc.subjectRedes neurais (Computação)
dc.subjectAnálise de séries temporais
dc.subject.capesEngenharias - Engenharia Elétrica
dc.subject.cnpqEngenharias - Engenharia Elétrica
dc.subject.keywordsTime series segmentation
dc.subject.keywordsStationarity analysis
dc.subject.keywordsTime series prediction improvement
dc.subject.keywordsSize reduction in time series
dc.subject.keywordsUnsupervised tuning of drift detectors
dc.subject.keywordsConcept drifts
dc.subject.keywordsNeural networks (Computing)
dc.subject.keywordsTime series analysis
dc.titleAdvances in time series analysis: segmentation, prediction, and unsupervised tuning of drift detectors
dc.title.alternativeAdvances in time series analysis: segmentation, prediction, and unsupervised tuning of drift detectors
dc.typeTese
dcterms.educationLevelDoutorado
dcterms.provenanceCentro de Tecnologia e Urbanismo

Arquivos

Pacote Original
Agora exibindo 1 - 2 de 2
Imagem de Miniatura
Nome:
EN_EEL_Dr_2024_Silva_Ricardo_P.pdf
Tamanho:
828.77 KB
Formato:
Adobe Portable Document Format
Descrição:
Texto completo ID. 196071
Baixar
Nenhuma Miniatura disponível
Nome:
EN_EEL_Dr_2024_Silva_Ricardo_P_Termo.pdf
Tamanho:
173.96 KB
Formato:
Adobe Portable Document Format
Descrição:
Termo de autorização
Baixar
Licença do Pacote
Agora exibindo 1 - 1 de 1
Nenhuma Miniatura disponível
Nome:
license.txt
Tamanho:
555 B
Formato:
Item-specific license agreed to upon submission
Descrição:
Baixar

Coleções

01 - Doutorado - Engenharia Elétrica