Metodologia para aprendizagem de ontologias: uma proposta no contexto do governo eletrônico

• Ano de defesa: 2025
• Autor(a) principal: Patrícia Lopes
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais
• Palavras-chave em Português: Ciência da informação; Ontologias; Governo eletrônico; Organização do conhecimento; Interoperabilidade semântica
• Link de acesso: https://hdl.handle.net/1843/83998

Ontologies have gained increasing recognition across various domains, being considered fundamental for the organization and representation of knowledge. However, the exponential growth of unstructured data on the Web has made the ontology construction process highly complex, time-consuming, and costly. In the context of e-government, this complexity is heightened by the need to structure and integrate large volumes of data in a standardized, transparent, and semantically interoperable manner. This dissertation aimed to develop a methodology for ontology learning focused on the organization and representation of knowledge and information within the Brazilian e-government environment. To this end, the following steps were carried out: analysis of studies on automatic and semi-automatic ontology generation; identification of Knowledge Organization Systems adopted by countries ranked highest in the United Nations E-Government Development Index; inventory of ontologies developed within the Brazilian government context; and comparative analysis of four ontology construction methodologies developed within the Graduate Program in Knowledge Management and Organization at the Federal University of Minas Gerais. As a main result, the Onto Merge Methodology (OMM) proposed a systematic and modular approach that enables partial automation of tasks related to knowledge acquisition and extraction, using techniques such as text mining and natural language processing. The OMM was validated through a proof of concept applied to the domain of public procurement, demonstrating efficiency gains and the potential for replication in other governmental contexts. It is concluded that, although full automation of ontology construction still faces technical and institutional challenges, methodologies such as the OMM represent a significant advancement by offering a more agile, replicable approach aligned with the scalability, standardization, and interoperability demands of digital public management.