Reducing Semantic Bias of Annotations for Semantic Web Service Discovery

摘要

Semantic web service discovery is very important in the whole semantic web service life cycle.Most of the current discovery methods are based on the stateless matching of declarative data semantics of service input and output parameters.The parameters of different web services which are annotated with the same concepts may represent instances with different semantics in the real world,i.e., there exists semantic bias between the semantics of concept and the semantics of the annotated parameter.This kind of bias seriously affects the discovery effectiveness.This paper proposes a method to reduce the semantic bias of annotations by adding some restrictions that are dynamically extracted from annotated context, and the resulting annotations are represented by the refined concept expressions.A novel semantic similarity measure for concept expressions is also proposed.Experimental results show that reducing the semantic bias dynamically can enrich the semantic web services and also improve the discovery effectiveness.In addition to its high effectiveness, the novel semantic matchmaker also shows its low computational complexity,which means it is suitable for practical application.