Community Detection in Collaborative Tagging Systems - Community-Built Databases: Research and Development - page 107

Chapter 5
Community Detection in Collaborative
Tagging Systems
Symeon Papadopoulos, Athena Vakali, and Yiannis Kompatsiaris
Abstract Collaborative Tagging Systems have seen significant success in recent
years as a convenient mechanism for organizing and sharing the favorite content of
web users. The collective tagging activities of users can be represented in the form
of a folksonomy, i.e., a tripartite network associating the users with the online
content resources of their selection and the tags used to annotate them. The network
structure of folksonomies has been extensively studied and exploited in a series of
information retrieval systems.
This chapter discusses the application of community detection, i.e., the identifi-
cation of groups of nodes in a network that are more densely connected to each
other than to the rest of the network, on folksonomy networks. In addition, we
describe a parameter-free extension of an existing community detection scheme
(Xu et al.: SCAN: A structural clustering algorithm for networks. In: Proceedings of
KDD'07: 13th ACM SIGKDD International Conference on Knowledge Discovery
and Data Mining, pp. 824-833. ACM, New York, NY (2007)) that is particularly
suited for discovering communities on tag networks, i.e., networks comprising tags
as nodes and associations among tags as edges. We found that the resulting tag
communities correspond to meaningful topic areas, which may be used in the
context of content retrieval and recommendation systems. The chapter discussion
is complemented bya set of evaluation tests on real tagging systems that demon-
strate that the proposed method produces more relevant tag communities than the
ones discovered by a state-of-the-art modularity maximization method (Clauset
et al.: Phys. Rev. E 70:066111, 2004).