Modeling interest graph of social networks with user-generated tags

Coming together with the increasing scale of social networks is the serious information overload problem, which makes it more and more necessary to conduct relevant researches on interest graph. To this end, we propose a process for modeling the interest graph of social networks with user-generated tags in this paper. According to the processing sequence over the tag data, the modeling procedure can be divided into three sub steps: firstly, using the Two-Mode network to One-Mode network mapping approach to construct an initial interest network represented by user tags; secondly, clustering the interest network to generate a hierarchal interest model by applying an improved GN algorithm; lastly, combining the social graph and interest graph to build an interest-based matching graph. Meanwhile, this paper also briefly describes the great potential application value of the interest graph in social networks.