Publikation
OWLS-MX3: An Adaptive Hybrid Semantic Service Matchmaker for OWL-S
Matthias Klusch; Patrick Kapahnke
In: Proceedings of 3rd International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web. International Workshop on Service Matchmaking and Resource Retrieval in the Semantic Web (SMR2-2009), 3rd, located at ISWC, October 25, Washington, D.C. USA, CEUR, Vol. 525, CEUR-WS.org, 2009.
Zusammenfassung
We present OWLS-MX3, the first adaptive hybrid semantic
Web service matchmaker for OWL-S. It learns how to best combine
logic-based, text similarity and ontological structure matching for hybrid
semantic selection of OWL-S services to given queries. For this purpose,
the matchmaker utilizes a SVM-based classifier which is learned over a
training set of the test collection OWLS-TC3 off-line. In particular, it
performs structural semantic matching to compensate for certain cases
of text matching failures caused by the observed characteristic of many
Semantic Web ontologies today of being rather mere is-a ontologies. Our
comparative retrieval performance evaluation experiments based on standard
measures for both binary and graded service relevance revealed
a rather negative result: There is a slight but not significant improvement
of performance over the non-adaptive variant OWLS-MX2 yet. On
the other hand, its adaptation feature clearly renders OWLS-MX3, in
principle, independent from any OWL-S test collection with referenced
ontologies and any kind of matching filters that are to be properly (re-
)combined by hand otherwise to reflect changes in the environment.