A comparison study of Flickr´s folksonomies and ontologies

Flickr is a widely-known Web 2.0 photo sharing system that allows users to describe their own photos with tags (aka. folksonomies) for indexing purposes. However, its tag-based photo retrieval function is severely hampered by the inherent nature of folksonomies. As the first step moving towards the alleviation of the problem, this paper, therefore, investigates characterizations of folksonomies on Flickr through its all-time most popular tags. In particular, the study focuses on tag clusters, related tags, and tag categorizations of the all-time most popular tags and compares them with concepts formally defined in ontologies from the Semantic Web. The experimental results imply the feasibilit of employing the Semantic Web technology in searching Flickr photos.