Binary Information Press
With the popularity of social tagging systems, research on folksonomy has attracted vast attention. Tag semantic ambiguity problem in folksonomy is one of the bottlenecks of the social tagging systems. Till now, most researchers try to identify the meaning of tags with the explicit or implicit precondition that each tag has only one meaning. Clearly, such methods can hardly avoid semantic information loss. In this paper, a novel tag semantic analysis paradigm by utilizing non-exclusive complex network clustering algorithm is proposed to tackle the tag semantic ambiguity problem with less semantic information loss.