Anonymous folksonomies for small enterprise webs: a case study

Tags and emergent folksonomies are a potentially rich new source of document annotations, offering query independent and dependent evidence for exploitation by information retrieval systems. Previous research has shown that tags may facilitate improved web search in an environment where each tagging action generates a (user, tag, resource) triple. For websites operated by a public institution, operational or privacy concerns may prevent the recording of data capable of identifying individuals. This leads to a simpler anonymous tagging system but is likely to reduce user motivation for tagging, since the user cannot access their own set of tags. It also means that votes for tags are not counted, and a potentially useful joining attribute is not available. Using webpage, metadata, query, click, anchortext and tag data provided by a public museum, we demonstrate that, despite these limitations, tag data collected by an anonymous tagging system has the potential to improve retrieval effectiveness.