Sistematização do processo de criação de definições formais em ontologias biomédicas: uma investigação no domínio das leucemias mieloides agudas

• Ano de defesa: 2015
• Autor(a) principal: Amanda Damasceno de Souza
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Definições formais; Ontologias Biomédicas; Sistemas de Organização do Conhecimento; Leucemia; Ciência da Informação; Ontologias (Recuperação da informação); Biomedicina
• Link de acesso: http://hdl.handle.net/1843/BUBD-AANE6D

Knowledge domains like cancer require technological solutions that make possible to organize and to deal with the volume of data, which exponentially increases insofar as research about the disease are developed. Ontologies are an alternative to the creation of models of reality in Biomedicine. The present work approaches the ontological theories applied to the construction of knowledge organization systems in biomedicine. In this context, a relevant issue is the lack of a methodic way of creating definitions for formal ontologies. The goal of this work is to established general principles to systematize the process of creating definitions suitable to be used in biomedical ontologies. In order to accomplish this, we developed part of an ontology in the domain of human blood; we created a list of basic principles that make possible a methodic creation of definitions; and, finally, we tested our basic principles in defining real biomedical entities related to human blood, specifically about leukemia. Our findings suggest that definitions for biomedical entities are generally created using a myriad criteria, which not always meet the good practices of building ontologies. After the test of our list of principles in real biomedical entities, we realized that some principles should be reviewed and improved in order that they can be properly applied in other sorts of biomedical ontologies, as well as other ontologies of the cancer domain. The main issues found in defining leukemia were circularity and lack of tangibility. The inherent complexity in diagnosing leukemia hampered the seek for the essence of classes. Creating definitions within the leukemia domain proved to be a hard and time-consuming task, which needs the participation of domain experts. We hope we can contribute in approaching open issues found in research fields so different as Biomedicine, Semantic Web, Librarianship, and Information Science.