Uma ontologia para o domínio de trabalhos de conclusão de curso de pós-graduação stricto sensu no brasil: implementação em dados ligados interoperáveis na web semântica

Abstract

The increase, in quantity and variety, of content on the Web has caused difficulties in terms of information retrieval (IR) that meets the needs of those interested in such information. The organization of knowledge is weakened by the absence of domain ontologies that adequately represent reality and the underuse of data on the Web best practices, such as the FAIR principles, especially with regard to interoperability, and this has also caused problems in IR. In this context, the semantic Web can offer publishing resources to allow IR through integrated queries with other databases and data repositories. Illustrating this situation, of organization and publication of data on the Web, deficiencies are observed in the set of postgraduate course conclusion works in Brazil. Given the above, the research problem of this work is how to organize the metadata of course completion papers using ontologies so that it is possible to retrieve them in an interoperable way? It is believed that saving time, minimizing expenses, expanding transparency and improving the quality of data and services provided by public sectors are part of publicity and efficiency. Aligned with the research problem, the general objective is to create a domain reference ontology and map it to an operational ontology to represent postgraduate course conclusion works in the context of the semantic Web and based on good publishing practices. The work has a qualitative research approach, applied nature and case study procedures. It used the ontology construction methodology called SABiO, supported by the UFO foundation ontology and the ontology-oriented modeling language OntoUML. As results, the conceptual model of the case study through a domain reference ontology, the design model for the operational ontology, the operational ontology implemented in Turtle, a database linked RDF located in a repository of a database graph oriented, and some SPARQL queries as a proof of concept. It was observed that the linked data serialized in RDF met a good part of the precepts of good practices of Web publishing. As a continuation of the research, it is indicated the validation of the artifacts developed in a substantially larger database and the development of interfaces to offer end users possibilities of queries with good usability.