Databases Reference
In-Depth Information
Modularity.
Re-using a part of an ontology requires to identify that part of the ontology
that does not 'interact' with the rest of the ontology when computing a given reasoning
task. Furthermore, it should be ensured that the importing ontology does not entail un-
wanted consequences w.r.t. the given reasoning task when importing the new ontology
module. These tasks are topics of the course material provided in [47] (with an emphasis
on tool support) and in [68,91] with an emphasis on the theoretical background.
Answering Conjunctive Queries.
As mentioned earlier, answering conjunctive queries
allows for a much more expressive query language than concept-based querying, i.e.,
instance queries. This reasoning task is currently a very active research area of DLs.
The DL Lite family of DLs is designed such that conjunctive query answering can be
performed efficiently. Query answering in DL Lite is covered in [26]. Methods for query
answering for both families of light-weight DLs is described in the course slides [32].
In addition to query answering for the light-weight DLs, [80] explains also the methods
for expressive DLs.
5
Application Areas of DLs
This section gives pointers introductory reading and course material on prominent ap-
plication areas for DLs.
Data integration.
Since DLs can represent information on different levels of detail they
are a good candidate for data integration. The core of the data integration approach
via DLs is their ability to capture other modeling languages frequently used to specify
database schemas—such as entity relationship diagrams (ER), see [40], module 2 or
UML, see [25].
Biomedical ontologies.
An account on the early use of DLs in the medical field is given
in [83]. In [100] an application in protein classifications described.
As a matter of F
A
CT, the medical ontology G
ALEN
[84] was the prime motivation
for the development of highly optimized reasoners [49]. Since then large biomedical
ontologies [94,34,85,92] have been valuable benchmarks for DL reasoners. More recent
overviews on medical ontologies written in DLs is given in [70,50].
Semantic Web.
The semantic web was early spotted as a potential application area of
DLs, since they allow to write ontologies in order to annotate web resources, see [7,45].
More importantly, the reasoning services defined and investigated for DLs support the
querying of these ontologies. From the plethora of course material on DLs and the
semantic web, we recommend to the reader [7,45] for early views on the subject and
[54,53] for more recent and technical ones.
The potential application of the semantic web facilitated the standardization of DLs.
An introduction to the original OWL standard can be found in [54] and OWL 2 and its
profiles is covered in [63,44,97].
