Modèle d'acquisition de données du LOD et génération automatique de requêtes
In
OD 2016, vol. RNTI-E-32, pp.79-110
Abstract
Nowadays, LOD (Linked Open Data) represents a promising source for many applications
of the Semantic Web. However, appropriate acquisition techniques have to be developed. The
work presented in this paper addresses this problem. The goal is to populate an OWL ontology
with property assertions taking into account the existence of multiple, equivalent and multivalued
properties in the LOD, as well as properties without values whose valuation might
be calculated elsewhere. Since the correspondences to establish are complex, we propose a
model to specify them. We also define a model to specify alternative paths to access properties
in case of missing values. Finally, we offer mechanisms to treat the acquired values. Secondly,
we show how these models are used to automatically generate SPARQL queries and thus,
facilitate interrogation. Two running examples are given.