Semantic Web service (SWS) discovery has gained more and more attention, leading to a great number of service matchmaking approaches. Existing approaches are based on SWS descriptions expressed according to a single specification (e.g., OWL-S, WSMO and SAWSDL). In this paper we propose a service matchmaking algorithm based on a SWS meta-model that abstracts the features of all the most common SWS specifications. The algorithm performs SWS comparison by increasingly relaxing matchmaking constraints, in order to maximize effectiveness of the discovery procedure, in terms of precision and recall. Moreover, to speed up algorithm performances, we provide SeeVa, an efficient representation of the SWS meta-model on which the algorithm is based. SeeVa is a storage system that includes a Datalog engine to enable language-independent reasoning capabilities. We evaluate the algorithm on public datasets containing SWS descriptions expressed using different specifications. Experiments demonstrate how the proposed approach outperforms main existing service matchmaking solutions both in terms of precision and recall and in terms of response time, thanks to the storage system and the Datalog engine.
SeeVa: A Model based framework for Semantic Web Service Discovery
BIANCHINI, Devis
2014-01-01
Abstract
Semantic Web service (SWS) discovery has gained more and more attention, leading to a great number of service matchmaking approaches. Existing approaches are based on SWS descriptions expressed according to a single specification (e.g., OWL-S, WSMO and SAWSDL). In this paper we propose a service matchmaking algorithm based on a SWS meta-model that abstracts the features of all the most common SWS specifications. The algorithm performs SWS comparison by increasingly relaxing matchmaking constraints, in order to maximize effectiveness of the discovery procedure, in terms of precision and recall. Moreover, to speed up algorithm performances, we provide SeeVa, an efficient representation of the SWS meta-model on which the algorithm is based. SeeVa is a storage system that includes a Datalog engine to enable language-independent reasoning capabilities. We evaluate the algorithm on public datasets containing SWS descriptions expressed using different specifications. Experiments demonstrate how the proposed approach outperforms main existing service matchmaking solutions both in terms of precision and recall and in terms of response time, thanks to the storage system and the Datalog engine.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
