Analysis of the information retrieval process in databases published as linked open data using the RDB2LOD approach

Abstract

Linked Open Data has become a standard for data publishing and data enrichment, and it supports the transition from a document-driven Web to an interconnected Web of data and thus to the Semantic Web. On the other hand, relational databases make up the core of most information systems currently in operation due to their maturity and efficiency in the form of storing and querying data. Thus, publishing the vast amount of data maintained in relational databases around the world in line with the good practices and recommendations of Linked Data can contribute significantly to the widespread adoption of Semantic Web tools and technologies. It is in this context that appeared the RDB2LOD approach for publishing Linked Open Data obtained from relational databases. However, once data is effectively published, the next step is efficiently and accurately searching and retrieving it for suitable use. This qualitative and exploratory work aims to analyze the information retrieval process in the RDB2LOD approach, in order to find out if the use of this approach can help to formulate SPARQL queries and, consequently, to improve the visualization and exploration of the retrieved data. For this, a bibliographic and documentary study was carried out, along with an experiment where the RDB2LOD approach's information retrieval process was evaluated in two different cases it was demonstrated that the consideration of the semantic aspects of terms in query expressions and the application of ontologies might improve data retrieval efficiency and accuracy.