A cluster-based approach for search and exploration of tag spaces

Although Semantic Web technology is increasingly becoming more and more important, tagging remains a popular method to describe Web resources. Therefore it is important to address the issues that are found in current tagging search engines, such as Flickr. We find that the free nature of tagging results in many issues for tag search engines, such as synonyms, homonyms, syntactic variations, etc. The Semantic Tag Clustering Search (STCS) framework deals with these issues by detecting syntactic variations of tags and by clustering semantically related tags. We evaluate our framework using Flickr data from 2009 and compare the STCS framework to two previously introduced tag clustering techniques.