The process designing domain ontologies from scratch is very time-consuming and is associatedwith a lot of effort. In the most cases, domain experts have defined XML Schemas, describingdomain data models, before ontologies have been created. Our idea is to generate ontologies outof XML Schemas automatically using XSLT transformations in a first step, and to derive domainontologies semi-automatically using SWRL rules in a second step. We apply our approach inorder to reuse the information located in the XML Schemas for the design of domain ontologies.In this paper, we aim to verify the hypothesis, that the effort and the time delivering high qualitydomain ontologies using the developed semi-automatic approach is much less than creating domainontologies in a completely manual way. We have applied the individual stages of the suggestedapproach to multiple different data models in the academic and the industry domain. Inaddition to that, we show one complete use case for which the traditional approach designingdomain ontologies manually and the proposed approach have been applied – the DDI-RDF DiscoveryVocabulary, which is an ontology of the social science metadata standard Data DocumentationInitiative.
展开▼