Feasability test: Integration of context-sensitive SPARQL query to DBPedia
Background: One central goal of the WDCOSF vocabulary broker is to virtually integrate research data from different scientific projects - IUGONET, ESPAS-FP7, ISDC - and possibly others in the future in a scalable manner by means of the Linked Data paradigm. The basic idea behind our system is to establish one conceptual "semantic core" - in our case hosted on the WDCOSF server - where the domain knowledge in terms of domain ontologies and controlled metdata vocabularies are combined and to integrate information pieces from the distributed research data resources on-the-fly by queries to appropriate SPARQL-endpoints.
In a first step we already interlinked three SKOS-ified metadata vocabularies - SPASE keywords, ESPAS keywords and GCMD Science keywords used by the respective projects by mapping the vocabulary concepts onto each other via <skos:closeMatch> and <skos:relatedMatch> relations. These vocabularies are locally stored in a Virtuoso triplestore on the WDCOSF server, and constitute the current base of our "semantic core".
Current step forward: As a preliminary test of the feasability to dynamically integrate external resources we added a context-sensitive SPARQL query to DBPedia, with the result that in the concept table an additional row shows up whenever the corresponding concept is found on DBPedia. Please note that this is only a first experimental step towards the integration of external research data resources into WDCOSF-System via SPARQL endpoints. Work is under way to integrate more relevant domain-specific resources.