The growing popularity of social tagging systems promises to alleviate the knowledge bottleneck that slows the full materialization of the Semantic Web, as these systems are cheap, extendable, scalable and respond quickly to user needs. However, for the sake of knowledge workflow, one needs to find a compromise between the ungoverned nature of folksonomies and the controlled vocabulary of domain-experts. In this paper, we address this concern by first devising a method that automatically combines folksonomies with domain-expert ontologies resulting in an enriched folksonomy. We then introduce a new algorithm based on frequent itemsets mining that efficiently learns an ontology over the concepts present in the enriched folksonomy. Moreover, we propose a new benchmark for ontology evaluation, which is used in the context of information finding, since this is one of the leading motivations for using ontologies in social tagging systems, to quantitatively assess our method. We conduct experiments on real data and empirically show the effectiveness of our approach.
Y. Chen, H. Dong, and W. Wang. Proceedings of the 2018 International Conference on Data Science and Information Technology, page 138--143. New York, NY, USA, ACM, (2018)
C. Wagner, P. Singer, M. Strohmaier, and B. Huberman. Proceedings of the 23rd International Conference on World Wide Web, page 735--746. Republic and Canton of Geneva, Switzerland, International World Wide Web Conferences Steering Committee, (2014)
W. Stock, and I. Peters. Joining Research and Practice: Social Computing and Information Science. American Society for Information Science and Technology, Milwaukee, Wisconsin, USA, (2007)