Effective browsing of personal Tag space in social tagging systems

Abstract

Social tagging systems such as del.icio.us and flickr have recently increased in popularity. The easy use of tags, along with the ability to publicly share tags and resources from others, has attracted many users to actively participate. However, more and more users are facing the problem of effectively organizing and browsing the increasing tags and resources in one’s personal tag space. In this paper, we present a novel approach to automatically group tags in one’s personal tag space into tag clusters thus helping the user organize the large tagged resources effectively. The proposed approach is based on the theory of concept lattices, which provides a powerful, well-founded, and computationally-tractable framework to model one’s personal tag space in which tags and collocated resources are represented and to compute such a transformation. We also propose a visual interface model for tag browsing by incorporating Tag Cloud and clustering technology, where tags in each cluster are organized as a tag-cloud. Experiments on del.icio.us show that the clustering results are very similar to bundles manually created by users. The presented algorithm not only groups the semantic-related tags into clusters, but also generates meaningful descriptions for each cluster automatically. In addition, the prototype system shows that the grouped Tag Cloud visualization interface improves the visual representation of a typical Tag Cloud layout.