A hypergraph model of social tagging networks

Abstract

The past few years have witnessed the great success of a new family of paradigms, so-called folksonomy, which allows users to freely associate tags to resources and efficiently manage them. In order to uncover the underlying structures and user behaviors in folksonomy, in this paper, we propose an evolutionary hypergrah model to explain the emerging statistical properties. The present model introduces a novel mechanism that one can not only assign tags to resources, but also retrieve resources via collaborative tags. We then compare the model with a real-world dataset: Del.icio.us. Indeed, the present model shows considerable agreement with the empirical data in following aspects: power-law hyperdegree distributions, negtive correlation between clustering coefficients and hyperdegrees, and small average distances. Furthermore, the model indicates that most tagging behaviors are motivated by labeling tags to resources, and tags play a significant role in effectively retrieving interesting resources and making acquaintance with congenial friends. The proposed model may shed some light on the in-depth understanding of the structure and function of folksonomy.

Modeling Consensus Semantics in Social Tagging Systems

“…The network was represented by a graph where edges run between items and users passing through tags with weights indicating tags given by users to items. Similar hypergraph model has also been used by Zlatic et al in [8], and Zhang et al in [9]. Peter Mika provided a similar semantic-social network model in [10] named Actor-Concept-Instance ontology model.…”

Proposing Ties in a Dense Hypergraph of Academics

“…Similar to the work here, given the use of NEs as features in the network, is a method of extracting names from websites to mine communities of people [27]. Some work on layered social networks implicitly adopts the hypergraph representation [25,26,44], though most social network analysis still relies on single-mode, unipartite graphs. What's more, most research using hypergraphs is framed as n-partite graph analysis, typically with a small number of layers [4,24].…”

From Popularity to Personality — A Heuristic Music Recommendation Method for Niche Market

“…Recently, some related studies have illustrated that using tagging information would benefit recommender systems on such a problem. Many researchers have discussed the problem from the perspective of both structure [19] and its applications in recommender systems [20][21][22] . Social tags can effectively build up relations between existing objects and the new ones.…”