A Novel Framework for Semantic Discovery of Web Services using Integrated Semantic Model

Semantic web technology plays a very critical role in the automatic web service discovery by assigning formal semantics to the service descriptions. Practically, it is not feasible to explicitly annotate the formal semantics to millions of existing services. Further, in user context, the request formation for services in semantic web is a complex process as it requires the user to be technically aware of the underlying technologies of the web services, discovery frameworks, description languages and implementation details. In this paper, we propose a semantic framework that enables Web service discovery based on the combination of semantic and syntax information contained in the service profiles. This novel approach for automatic discovery of Web services employs measures of semantic relatedness, Natural Language Processing techniques and information retrieval based statistical models to match a user request. Additionally, we present an efficient semantic matching technique to compute the intra service semantic similarity scores which further facilitates semantic ranking of services. The efficiency of the proposed approach has been demonstrated through experimental evaluations which clearly show that high degree of automation can be achieved with high precision. The results have been further authenticated by providing comparisons with other Information Retrieval based methods.