From software APIs to web service ontologies: a semi-automatic extraction method

Successful employment of semantic web services depends on the availability of high quality ontologies to describe the domains of these services. As always, building such ontologies is difficult and costly, thus hampering web service deployment. Our hypothesis is that since the functionality offered by a web service is reflected by the underlying software, domain ontologies could be built by analyzing the documentation of that software. We verify this hypothesis in the domain of RDF ontology storage tools.We implemented and fine-tuned a semi-automatic method to extract domain ontologies from software documentation. The quality of the extracted ontologies was verified against a high quality hand-built ontology of the same domain. Despite the low linguistic quality of the corpus, our method allows extracting a considerable amount of information for a domain ontology