Database Reference
In-Depth Information
9.3.14 P
rOPerty
c
hains
anD
g
eneraL
c
OncePt
i
ncLusiOn
a
xiOms
Another useful feature introduced in OWL 2 is the property chain, which is also
known (for certain reasoners and so you may come across this term) as a
complex
role inclusion axiom
. A property chain defines a property as a chain of object prop-
erties. The clearest example of this is in the domain of families: An aunt is the sister
of one's parent. So, the property
hasAunt
can be defined as the property chain
hasParent o hasSister
(the DL symbol
o
is often used to denote a link in a
property chain).
Everything that has a Parent that
has a Sister will also have that
as an Aunt.
Everything that has a Part that
contains some Thing will also
contain that Thing.
ObjectProperty: hasAunt
subPropertyChain: hasParent o
hasSister
ObjectProperty: contains
subPropertyChain: hasPart o
contains
You may also come across the term
general concept inclusion
(GCI) axiom,
which is a more specific version of the complex role inclusion axiom, involving a
concept associated with the property chain, for example, “Everything that has a Part
that contains some Water will also contain some Water”.
A GCI occurs when C is a subclass of D, where C is any general concept and C
and D can be complex (e.g., anonymous classes). A GCI can be thought of as any
statement that has more than just a single class name as the subject. The more of
them there are in your ontology, the longer it will take to reason over. This is also true
for the property chains. However, these types of axiom are often the “glue” that holds
the ontology together and helps you achieve the right answers to your competency
questions. If, when you query your ontology with the competency question “test set,”
you do not see the answer you are expecting, consider whether you need to add some
complex role inclusion axioms or GCIs to create the correct chain of reasoning.
9.4 TOOLS FOR AUTHORING
To manage the creation and maintenance of an ontology, tool support is essential.
There are two main types of ontology tool—editors and reasoners—although some
products combine the functions of both. An ontology editor is used to create and edit
an ontology, while a reasoner allows queries to be asked of the ontology and deter-
mines the implicit knowledge that is a consequence of the ontology's statements.
The most widely known tool for creating OWL ontologies, as well as RDF/XML,
is the free, open source editor Protégé,
8
produced at Stanford University (Musen,
1988). Protégé 3.x is a frames-based editor, that is, one that organizes an ontology
into a set of classes, slots (for properties and relationships), and instances and has
been evolving over a number of years. The more commonly used version is Protégé
OWL (Protégé version 4), an extension of the Protégé editor built at the University of
Manchester. This allows OWL and RDF ontologies to be loaded, saved, edited, and
visualized. In addition, rules encoded in the Semantic Web Rule Language (SWRL)
(Horrocks et al., 2004) can be added. There are several useful plug-ins, for example,
