[Back]

A. Mazak, M. Lanzenberger, B. Schandl:
"iweightings: Enhancing Structure-based Ontology Alignment by Enriching Models with Importance Weighting";
Talk: 4th International Conference on Complex, Intelligent and Software Intensive Systems (CISIS 2010), Krakow, Poland; 2010-02-15 - 2010-02-18; in: "Proc. of CISIS 2010", L. Barolli et al. (ed.); IEEE Computer Society, (2010), ISBN: 978-0-7695-3967-6; 992 - 997.

Structural ontology matching methods analyze
mainly two factors: entity labels and relationships among
entities. We propose to additionally consider an importance
and relevance factor, which is determined by two indicators
automatically calculated by a (simple) weighting method.
This weighting factor represents the importance of a concept
based on its information significance in the modeling
context and, additionally, its relevance for structure-based
alignment depending on the number of relationships this
concept participates in quantified by the rweighting indicator.
The method starts via a manually weighting annotation of
relationships among concepts conducted by ontology engineers
during the ontology development process. Our approach is
an assistance mechanism to improve the ontology alignment
process and to enhance the cognitive support for users. Thus,
ontology alignment becomes already important ex ante when
the ontology development process starts, unlike other alignment
techniques, which consider only ex post knowledge.

Ontology alignment, Meta-Object Facility, ECore meta-model, importance and relevance weighting, candidate ranking.

http://dx.doi.org/10.1109/CISIS.2010.182

http://publik.tuwien.ac.at/files/PubDat_187532.pdf