Extração de conceitos e relações taxonômicas usando análise de conceitos formais e agrupamento fuzzy de dados

• Ano de defesa: 2017
• Autor(a) principal: Lima, Suzane Carol de
• Orientador(a): Camargo, Heloisa de Arruda
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• Palavras-chave em Português: Reticulado de conceitos; Redução; Agrupamento Fuzzy; Concept lattice; FCA; Reduce; Fuzzy clustering
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/12011

Some structures for knowledge representation are organized from concepts and relationships between concepts, among which we can mention semantic networks and ontologies. An important tool that help in the creation process of these structures is the Formal Concept Analysis (FCA). FCA has been applied in several fields of research, such as data mining, machine learning, artificial intelligence and Software Engineering. The FCA can now be considered an important formalism for the representation of knowledge, extraction and analysis with applications in diferente areas, and is used for the construction of ontologies, since it provides a basis for the development and implementation of methods to extract ontological concepts as well as the ontological taxonomy involving the extracted concepts. In the Formal Concept Analysis, concepts are sets of objects that share the same attributes. Concepts are extracted from a set of data and organized in the form of a Concept Lattice, defined by the relation of inclusion between concepts. The structure of the Conceptual Framework can become large due to the high number of concepts and relations, making a complex structure, and often difficult computational process. The purpose of this work is to reduce the formal context of a specific domain by using two fuzzy clustering algorithms, so that a reduced Concept Lattice is generated. The results showed that the Fuzzy C-Means clustering algorithm performed better than Possibilistic Fuzzy C-Means algorithm.