Automatically identifying bounds on semantic annotations for bioinformatics web service input parameters

Abstract

Semantic annotation of web services is achieved by associating parts of service interface descriptions to concepts defined in a semantic model, such as an ontology. Annotating web services with semantic metadata is currently a manual procedure that is conducted by human annotators. In practice, however, service annotation can be both time-consuming and error-prone to implement on real-world web services. We have investigated this problem in the bioinformatics domain, where certain characteristics shared by many web services make them a difficult breed to annotate with semantics. By taking these characteristics into consideration, we present a probing-based technique that identifies upper and lower semantic bounds for the annotation of web service input parameters. The resulting bounds can serve as a useful starting point for human annotators and could help reduce the required effort in the semantic annotation process. © 2010 Springer-Verlag.