FOCA: uma metodologia que utiliza princípios da representação do conhecimento para avaliação de ontologias

• Ano de defesa: 2015
• Autor(a) principal: Bandeira, Judson Melo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Alagoas; Brasil; Programa de Pós-Graduação em Modelagem Computacional de Conhecimento; UFAL
• Palavras-chave em Português: Ontologia; Ambientes virtuais de aprendizagem; Ontology; Virtual learning environments; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://www.repositorio.ufal.br/handle/riufal/1731

Ontology is a term originated in Philosophy through the Greek terms "ontos" (being) and "logos" (study), which means the study of being qua being. This term was brought to Computer Science as a computational artifact that represents knowledge of a certain universe of discourse. One way to do this is through the five roles of knowledge representation: substitute, ontological commitment, intelligent reasoning, computational efficiency and human expression. In computer science, ontologies are applied in many areas such as Requirements Engineering, Database, Semantic Web and Linked Data. The subarea responsible for ontology building process is ontology engineering, which has about twenty methodologies. However, knowing well the domain, defining the concepts involved in it correctly and at the same time comply the knowledge representation roles make ontology modeling an arduous task. Thus, how bigger is the complexity of ontology, the more its quality tends to decrease, reducing the reuse’s potential. Consequently, the quality of the evaluation process of an ontology becomes a fundamental task. The literature of this area proposed different approaches for ontology evaluation, but they are based on quality metrics that do not guide the evaluators in decision making about the quality of ontology. As a result, the establishment of ontology quality becomes more dependent on the experience of the evaluator and less dependent on the knowledge representation criteria. One way to simplify this process is to guide the evaluator about what criteria are related to each role of knowledge representation. This work establishes a methodology for ontology evaluation, which is based on a match between the roles of knowledge representation with the main quality criteria for the ontology. This match is built by Goal, Question, Metric approach. This methodology consists of three phases: Ontology Type Verification; Questions Verification and ontology Quality Verification. This methodology was validated by conducting an experiment that aims to show that the methodology leads evaluators with different experiences in ontology to the same evaluation result. As a result of the experiment, was possible create a formula for global and partial scores for the quality of ontology, validating it.