Databases Reference
In-Depth Information
Query 4
SELECT ?X
WHERE { ?X a foaf:Agent .}
Clearly, for the RDF data in Fig. 2, this query would not return any solu-
tions (despite many implicit instances of the class being available), whereas:
Query 5
SELECT ?X
WHERE { ?X a foaf:Person .}
would return all the company founders listed in DBpedia, since these are
explicitly typed as foaf:Person s .
We emphasise that these queries miss (implicit) results that would be intuitively
expected as an answer and that could be achieved through reasoning. In the
following we will substantiate this intuition referring to some of the challenges
(C1-C5) mentioned in the introduction; before that, however, we need to clarify
the importance of schema information and its semantics in RDF.
2.4
Inferring Additional Triples by Schema Information and Rules
In order to model schema information, which also allows to infer additional
implicit information in RDF, the Semantic Web offers two ontology languages:
the lightweight RDF Schema (RDFS) [13] and the expressive Web Ontology
Language (OWL) [37]. Within this section, we will briefly cover an overview of
the essential features of these languages in a Linked Data setting.
RDFS RDF itself already provides means to express class membership
( rdf:type ) ; RDF Schema (RDFS) additionally provides a special vocabulary,
consisting primarily of RDF properties with a predefined semantics to
model class hierarchies ( rdfs:subClassOf ) , and property hierarchies
( rdfs:subPropertyOf ) , as well as means to define domains and ranges that
respectively allow for associating a class to the subject and object of a relation
with a given property ( rdfs:domain , rdfs:range ) . These core RDFS properties
allow for describing and embedding the semantics of user-defined vocabularies
in RDF itself.
OWL The Web Ontology Language (OWL) extends RDFS and allows for ex-
pressing further schema definitions in RDF, e.g., allowing to express equal-
ity of individuals ( owl:sameAs ) , equivalence or disjointness of properties and
classes ( owl:equivalentClass , owl:equivalentProperty , owl:disjointWith ,
owl:propertyDisjointWith ) , or complex class definitions; while a full account
is beyond our scope, more details on additional OWL features will be discussed
in Section 4 below.
Search WWH ::




Custom Search