Estudo de alternativas de mapeamento de ontologias da linguagem OntoUML para OWL: abordagens para representação de informação temporal

Abstract

The application of ontologies for domain knowledge representation has gained much interest in the later years, specially in the context of Semantic Web. However, it is been observed a strong tendency a strong tendency to prioritize the computational efficiency over quality of representation, leaving out important features of the domain, such as temporal aspects. Indeed, the OWL, standard adopted for the Semantic Web, does not allow, in principle, the representation of temporal information. The main strategies proposed in the literature to solve this problem, namely, the Perdurantistic and Temporal Reification approaches, do not provide support for the modeling support the decisions, leaving the responsibility for the user. In this context, this research aims, through the application of results from the philosophical discipline of Formal Ontology, to provide for such approaches a base structure ontologically founded and guidelines to guide its use, evaluating and comparing the proposed alternatives. Moreover, since not a trivial task the direct application of the ontological notions, it is observed that using a modeling language that incorporates them, it is possible to define systematic mappings from this language to OWL so that it allows automatic generation of models in an Ontology Engineering process. Thus, another goal is to propose alternative mappings between the ontological level language OntoUML and epistemological level language OWL. Moreover, one of these alternatives have been applied in a case study in the Electrocardiography domain for automatically codifying an ECG reference ontology following an Ontology Engineering process, allowing to represent the temporal information. Finally, the results of this research are relevant in the context of Ontology Engineering for providing systematic mappings which, once implemented, allows the reference ontology codification to be conducted in an automated and efficient way, while preserving the quality of the reference model and avoiding errors arising from manual coding process and, moreover, allowing to circumvent the limitation of OWL on the representation of temporal information.