Unsupervised Fuzzy eIX: clusterização interna-externa fuzzy evolutiva de fluxos de dados não-estacionários

• Ano de defesa: 2022
• Autor(a) principal: Aguiar, Charles Carvalho de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Engenharia de Sistemas e Automação; UFLA; brasil; Departamento de Engenharia
• Palavras-chave em Português: Aprendizado não supervisionado; Sistema Fuzzy evolutivo; Computação granular; Fluxos de dados online; Unsupervised learning; Evolving Fuzzy system; Granular computing; Online data stream; Automação Eletrônica de Processos Elétricos e Industriais
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/49682

Classifiers with time-varying decision boundaries, namely, evolving classifiers, play an important role in a scenario in which information is available as an online data stream. This text presents a new unsupervised learning method for numerical data called evolving Internal-eXternal Fuzzy clustering method (Fuzzy eIX). The notion of double-boundary fuzzy granules and some of its implications are developed and explored. It will be shown how type 1 and type 2 fuzzy inference systems can be obtained from the projection of Fuzzy eIX granules on orthogonal axes corresponding to the dimensions of a problem. Fuzzy eIX learning algorithm performs Pedrycz Balanced Information Granularity principle within fuzzy eIX classifiers to achieve a higher level of model understandability in a given problem domain. Internal and external granules are updated from a numerical data stream at the same time that the global granular structure of the classifier is autonomously evolved. A synthetic preliminary problem called Rotation of Twin Gaussians shows the behavior of the classifier for a nonstationary data stream input. Additionally, the performance of the Fuzzy eIX method will be compared to other two evolving methods already established in the literature when it comes to the classification of benchmark data sets usually employed in online machine learning models assessments. Comparisons will also be conducted in terms of partition quality through incremental cluster validation indexes, the accuracy and compactness of the resulting rules structure.