Risk driven semantic webservice discovery
in unstructured P2P networks
Diploma Thesis by Ulrich Basters
University of the Saarland, Computer Science Dept./DFKI
RS2D is an approach to solve the problem of decentralized retrieval of semantic web services in unstructured P2P networks with reasonably high retrieval performance in terms of precision and low communication costs without any prior knowledge about the environment including services, ontologies, or network topology.
Key idea of RS2D is to allow each autonomous peer to dynamically build up and manage its own local ontology based on a minimal shared vocabulary only, and to quickly learn which of its direct neighbours in the network it knows about will probably return relevant semantic web services for a given query with minimal mixed Bayesian risk of both semantic loss and high communication in total.
Section 1 - RS2D V1.0 Evaluation
The results of the experimental evaluation of the original version of RS2D (version 1.0) are available here.
Basters, U.; Klusch, M. (2006): RS2D: Fast Adaptive Search for Semantic Web Services in Unstructured P2P Networks. Proceedings of the 5th International Semantic Web Conference (ISWC 2006), Athens, USA
Section 2 - RS2D V2.0 Evaluation
The results of the experimental evaluation of the second version of RS2D (version 2.0) which exploits a different semantic loss function are available here. The performance is drastically improved over v1.0 but at the cost of communication.
Basters, U.; Klusch, M. (2006): Risk Driven Semantic P2P Service Retrieval. Proceedings of the 6th IEEE International Conference on P2P Computing (P2P 2006), Cambridge, UK, IEEE CS Press.
Basters, U. (2006): RS2D - Risikobasierte semantische Suche nach OWL-S Diensten in unstrukturierten P2P-Netzen. Diploma Thesis, University of the Saarland, Computer Science Dept./DFKI
RS2D has been developed by
Ulrich Basters, Matthias Klusch
German Research Center for Artificial Intelligence (DFKI),
Saarbruecken, Germany, 2006
Contact: email@example.com, firstname.lastname@example.org