Quality Metrics for Tags of Broad Folksonomies

Folksonomies do not restrict its users to use a set of keywords preselected by a group of experts for interpersonal information retrieval. We assume that the quality of tag-based information retrieval and tag suggestions when tagging a resource can be ameliorated if we are able to automatically detect the tags that have an intersubjective meaning or tags that are understood and used by many members of a group. In this paper, (1) we suggest three possible tag quality measures for broad folksonomies (2) by means of an analysis of a del.icio.us dataset, we provide preliminary evidence that the suggested metrics return useful sets of intersubjectively valid tags and (3) as an additional evaluation, we asked individuals to judge the tag sets obtained through the metrics.