Um método para publicação semântica Linked Data de bases de dados convencionais e um estudo de caso real de artigos acadêmicos

Abstract

Using a quali-quantitative, applied, and experimental research methodology, this paper proposes a method for systematic mapping and publication of an existing relational database according to the Linked Data principles and based on a case study of academic papers of the internal conference Semana de Integração Acadêmica (Academic Integration Week), carried out in a Brazilian federal public university. The proposed method results from mapping the knowledge domain studied in reputed Linked Data ontologies (Schema.org, Friend of a Friend, Bibliographic Ontology, Semantic Web Conference Ontology, etc.) and it was applied to the conference’s relational database in order to make it available in machine readable format on the Web, establishing, in addition, semantic links with the famous DBpedia dataset through an automated mashup process. The results obtained with the method were quite satisfactory. The goal of publishing a Linked Data view over the relational data was fully achieved without changing it. With this work, we hope to foster making semantic data available on the web in accordance with Linked Data principles, thus contributing to a wide dissemination of knowledge, boosted by the ability the Semantic Web provides to machines for interconnecting, understanding, and discovering information