SAWSDL-MX2: A Machine-Learning Approach for Integrating Semantic Web Service Matchmaking Variants

Matthias Klusch, Patrick Kapahnke, Ingo Zinnikus

In: Ernesto Damiani, Rong Chang, Jia Zhang (editor). 2009 IEEE International Conference on Web Services. IEEE International Conference on Web Services (ICWS-2009) 7th July 6-10 Los Angeles CA United States Pages 335-342 ISBN 978-0-7695-3709-2 IEEE Press 2009.


In this paper, we present SAWSDL-MX2, a hybrid semantic Web service matchmaker for SAWSDL services. Building on our initial work, we adopt logic-based as well as text similarity service selection for model references and add a structural approach, which operates on the pure syntactic description of WSDL elements. The integration of these matching variants is accomplished using a Support Vector Machine (SVM) with non-linear kernel, thus automatically adapting an aggregation function based on previously experienced training data. Results of our performance evaluation based on the standard measures recall and precision over the SAWSDL-TC1 test collection as well as an exhaustive example for all basic matching variants are also given.

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz