A Domain-based Approach to Publish Data on the Web

The Linked Data project defines a set of practices for publishing structured data on the Web. In order to apply the recommended practices and achieve this Web of Data vision, existing data provided in diverse formats should be converted to a standard model. Since these data are in nature heterogeneous, it is usually unfeasible to convert them without considering the knowledge domain in which they exist. In this light, we propose an approach, named SenseRDF, which makes use of a domain reference, made available by vocabularies and domain ontologies, to provide the semantics during the conversion of a dataset from one format to a standard RDF model. Also, we adopt a semi-automated technique, where a domain expert is required to assist the conversion process, which is incrementally enriched. We present the principles underlying our approach, some usage examples and the obtained experimental results.