WordNet in RDF/OWL, Best Practice

February 20th, 2006

WordNet has always been an interesting source of knowledge. It began in the mid 80s as a project by Princeton’s George Miller with the aim of exploring new, computer-mediated models of dictionaries. I recall hearing him talk about this in 1984 or 1985.

WordNet is what might be called a lexical ontology, i.e. one focused on words and their meanings. It currently consists of about 150K words associated with 115K synsets (word senses or meanings) and is widely used as a resource for AI, NLP and IR projects. WordNet has been mapped into RDF many times and because of its broad coverage has been used in many ways, especially for annotating objects and ontology mapping. Using wordNet senses for tagging, for example, can provide more accuracy and reduce ambiguity.

The WordNet Task Force of the SWBPD WG has released a editor’s draft of RDF/OWL Representation of WordNet. It provides a standard conversion of WordNet for direct use by Semantic Web application developers. “By providing a standard conversion that is as complete as possible the TF aims to improve interoperability of SW applications that use WordNet and simplify the choice between the existing RDF/OWL versions.” (spotted on SWIG scratchpad)