| Summary: | Folksonomies have become very popular as means to organize large sets of resources shared over the Social Web. The bottom-up nature of folksonomies has proved to be an interesting alternative to the current effort at semantic web ontologies since folksonomies provide a rich terminology generated by large user-communities. Besides, ontologies extracted from folksonomies can represent the intelligence collective of social communities. Such ontologies also represent a core element of a new feature of the Web, the Internet of Things. Many research studies have captured semantics in folksonomies, some of which have developed ontologies from folksonomy. However, the formal specific-domain ontology consisting of domain-dependent relations has not been researched yet. This paper introduces an algorithm for deriving a domain-specific ontology from folksonomy tags. The proposed algorithm starts by collecting a domain-specific terminology; next, discovering a pre-defined set of conceptual relationships among the domain terminologies. The evaluation of the algorithm, using a dataset extracted from BibSonomy, demonstrated that the algorithm could effectively learn domain ontologies consisting of domain concepts linked by meaningful and high accurate relationships. Furthermore, the proposed algorithm can help reduce common issues related to tag ambiguity and synonymous tags.
|
|---|
