Descripción de fuentes de datos heterogéneas utilizando tecnologías semánticas

Abstract

The Semantic Web, propound potential opportunities to give meaning to web content. Ontologies are one of the main tools to explicitly specify the concepts of a particular domain, its properties and its relationships; so that the information is published in formats that are understandable by machine agents that can locate and manage the information accurately. This thesis presents an application for the generation of a common ontological model, which describes different data sources through its metadata, specifically worked with sources such as Database, CSV, XML and EXCEL. For this, different ontologies of metadata description were analyzed, among which DCAT, PHDD and DISCO are available. These three sources were united in a single model, on which some modifications were made, the most important being the incorporation of a structure, which allows me to describe the different types of data that have the attributes of the sources. Afterwards, a common relational model was created, where the extracted metadata is temporarily stored, for its subsequent mapping with the common ontological model. Finally, we proceeded to generate an RDF file on the common ontological model and publish the same for its exploitation. In order to experimentally validate the created model, an integration scenario was set up for several data sources, where a comparison was made by consulting the common ontological model and manually inspecting the metadata and data of said sources, this with the purpose to obtain the utility of the common ontological model. Concluding that through the common ontological model, the user in charge of integration can extract the possible semantic assignments in terms of data integration, easily and at a lower cost, since it is done on a single common repository.