Inteligent Tag Grouping By Using An Aglomerative Clustering Algorithm

Abstract

Tagging can be defined as a process of assigning short textual descriptions or key-words (called tags) to information objects. It is a simple approach to information organization that was regularly practiced over the last years. Tagging systems usually have relatively flat tags. This means that while one can easily browse by a tag, one cannot as easily see tags that have wider or more specific meaning than a given tag. It is also difficult to get a broad overview of the tags that do exist in the tagging systems, aside from frequency based displays like tag clouds. In this paper we investigate how correlated tags can be grouped by using an agglomerative clustering algorithm considering only the label part (output space) of the data. We have applied this approach on the StackOverflow tag cloud and discuss the obtained results.