100 likes | 237 Views
Towards Conceptual Structure Verification of Linked Data Vocabularies. Vojtěch Svátek, Miroslav Vacura, Martin Homola, and Ján Kl’uka. University of Economics , Prague. Comenius University in Bratislava. Linked Data. Adherence to publicly available and widely shared vocabularies
E N D
Towards Conceptual Structure Verification of LinkedData Vocabularies Vojtěch Svátek, Miroslav Vacura, Martin Homola, and Ján Kl’uka University ofEconomics, Prague Comenius University in Bratislava
Linked Data Adherence to publicly available and widely shared vocabularies • FOAF (for personal profiles) • Dublin Core (for bibliographic metadata) • Music Ontology(MO; for recordings and other musical information) • GoodRelations(GR; for e-commercedata).
Surface and deep model Theactual RDF/OWL representationoften only roughly approximates the structure of the implicit conceptual model that faithfully captures the real-world state of affairs. Surface model Deep model
Example ex:MyCompanygr:hasBusinessFunctiongr:Repair Surface model Individual Individual Individual Quality Deep model
Deep model structure • D-object refers to a real-world object, which can be tangible (such as people, animals,things, etc.) orintangible (variousabstractentities such as topics, processes, etc.). • D-class refers to a real-world class. Therefore it is usually a set of D-individualsinstantiating the same concept or sharing a common property. For simplicity, it alsocovers the notion of quality (e.g., red color).
Deep model structure • D-relationshiprefersto a (particular) relationship between two entities (for instance, an individual is producedby some producer, or a person owns some individual). • D-relation, which refers to a real world conceptual relation, a class of relationships
Deep model structure • D-valuation refers to an assignment of a data value to some entity (e.g.,some person has height of 199 cm). • D-attribute, referringto real world valuations of the same (usually quantitative) property.
Use-case - Music Ontology • To demonstrate a practical use of the deep model, we now have a look on the MusicOntology vocabulary with the purpose of ‘deep disambiguation’ of selected constructs. (In paper)
NextSteps • Evaluation- mapping for MO, GR, and FOAF, in the form of annotation of individualsurface entities (in particular, classes, individuals, and property ranges) with ‘meta-properties’ • Tool Support- Protégé plugin that would allow to createannotations • Practical Application- inspection of vocabularies and datasets (summaries) via inviewertransformation among different surface representations of the same deep structure