[DL] Ontology Completion Plugin for Protege 4

Baris Sertkaya sertkaya at tcs.inf.tu-dresden.de
Wed Feb 18 17:33:39 CET 2009



Dear list members,

I would like to announce a Protégé 4 plugin that I have developed for
ontology completion.

I have developed an ontology completion tool called *OntoComP*, which
stands for Ontology Completion Plugin. It is implemented in Java as a
plugin to the Protégé 4 ontology editor. The plugin, the source code and
the documentation on howto use it are available under:
http://ontocomp.googlecode.com

OntoComP enables the user to check whether an ontology contains 'all
relevant information' about the application domain, and to extend the
ontology appropriately if this is not the case. It achieves this by
asking the user questions of the form
"are instances of classes C1 and C2 also instances of the class C3?"
If the user confirms such a question, then a new axiom of the
application domain (that does not already follow from the ontology) has
been discovered, and it is added to the ontology. If she rejects it, she
is expected to give a counterexample to this question. In this case, a
new individual of the application domain that is not represented in the
ontology has been discovered, and it is added to the ontology. The
approach guarantees that on the one hand when all such questions are
answered, the ontology is complete w.r.t. a fixed interpretation, and,
on the other hand, the interaction with the user is kept to a minimum.

The approach is based on the extension of a well-known knowledge
acquisition method developed in Formal Concept Analysis, which was
presented in [1].

I would appreciate feedback and user comments on OntoComp, and would be
very happy to provide support in using it. New developers, code reviews
and other developer efforts are also very welcome.

Best regards,
Baris Sertkaya


-------------------
[1] F. Baader, B. Ganter, U. Sattler, and B. Sertkaya. Completing
Description Logic Knowledge Bases using Formal Concept Analysis. In
Proceedings of the Twentieth International Joint Conference on
Artificial Intelligence (IJCAI-07).
AAAI Press, 2007.







More information about the dl mailing list