Um estudo de caso sobre aquisição do conhecimento em ontologias: validação de conhecimento especializado a partir de linguagens naturais controladas

Detalhes bibliográficos
Ano de defesa: 2021
Autor(a) principal: Murillo Lima Modesto
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal de Minas Gerais
Brasil
ECI - ESCOLA DE CIENCIA DA INFORMAÇÃO
Programa de Pós-Graduação em Gestão e Organização do Conhecimento
UFMG
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
Link de acesso: http://hdl.handle.net/1843/36662
https://orcid.org/0000-0002-8344-4832
Resumo: Knowledge Acquisition is an important, but not studied, stage in the process of constructing ontologies. It constitutes a gap in the understanding by domain experts and even information professionals. Natural language is used in communication, it is informal and easily understood by humans. A formal language consists of the expression of natural language through a formal syntax, well defined in the sense of being restricted, with semantics free of ambiguities and supporting the formal methods of reasoning. Formal languages are adopted in the process of constructing ontologies, which are formal artifacts of Knowledge Representation. In the construction of these artifacts, after the Knowledge Acquisition and Conceptualization stages, one needs to validate the entire representation with domain experts, in order to confirm that the model really represents the specialized knowledge. However, the logical language employed in ontologies is not always easy for humans. In this context, a research issue, within the gap between the formal language of the ontological artifact and the natural language of the specialist, was identified based on the difficulty and even impossibility of proper validation of the ontology. Thus, we seek to find answers to the following question: how to plan and carry out the validation of formal representation of specialized knowledge in ontologies by domain experts by using an interface based on controlled natural language? Our general goal is to create a method to validate the acquisition of knowledge in comparing the results of that stage with the axioms of the formalization phase using controlled languages. The research is justified by the little literature found both on the methods of knowledge acquisition and on the validation of formal representations, even though both are relevant to the research of ontologies in Information Science (IC). To achieve this goal, a fragment of a domain ontology was taken from the Patrimonial Control Manual for the Electric Sector, developed by Brazilian National Electric Energy Agency. The axioms related to the chosen terms were translated using friendly language to represent ontologies, and after the translation, validated by a domain experts. Understanding how knowledge is extracted and registered by domain experts, consolidated by CI specialists and, finally, validated by domain experts impacts on the quality of the ontological artifacts produced and in their reuse, configuring the main contribution of this research.