Databases Reference
In-Depth Information
to RDFS/OWL techniques in that it allows for integrating numerical infor-
mation embedded in Linked Data in a manner that RDFS and OWL do
not.
We then briefly recap and conclude the lecture material in Section 8.
2 Preliminaries
As mentioned in the introduction, more and more Web sources provide readily
accessible and interlinked RDF data. In the context of this paper, when speaking
about structured data on the Web, we focus on data represented and published in
the Resource Description Framework (RDF) according to the Linked Data prin-
ciples. Hundreds of such datasets have been published in this manner, and have
been collectively illustrated by Cyganiak and Jentzsch in the so-called “Linked
Data Cloud”.
1
Among these DBpedia
2
plays a central role as an RDF extract
of structured data from Wikipedia. As a further example, the New York Times
3
provide RDF metadata about entities occurring in their articles as well as API
access to the latest articles about these entities.
In terms of reasoning, our main considerations revolve around deductive infer-
ences that enrich this RDF data with implicit information by means of exploit-
ing (parts of) the formal semantics of RDFS and OWL. In particular, we aim
at motivating how such additional inferences can contribute to query answering
using SPARQL queries over the whole body of Web-accessible Linked Data (per
Berners-Lee's “one big database” vision [9]).
We begin by introducing the basic concepts of the relevant Web standards
along with a running example from real-world Linked Data. As a sort of dis-
claimer upfront, we emphasise that these preliminaries are not intended to re-
place a fully-fledged introduction to RDF, SPARQL, OWL or Description Logics;
we refer the interested reader to excellent introductory chapters published within
the earlier editions of the Reasoning Web summer school or to the ocial current
W3C standard specifications in the references for further details.
2.1 The Resource Description Framework - RDF
Informally, all RDF data can be understood as a set of subject-predicate-object
triples, where all subjects and predicates are URIs, and in the object position
both URIs and literal values (such as numbers, strings, etc.) are allowed. Fur-
thermore, blank nodes can be used in the subject or object resource to denote
an unnamed resource with local scope. Some sample RDF data in the popular
Turtle syntax [5,6] are shown in Fig. 1.
