Graph-Based Hierarchical Categorization of Microblog Users

Microblogging has created a big social network with big social media data. Modeling and analyzing the relationships of behaviors among micro log users, and achieving the inherent categories or communities, are paid much attention in social network and big data paradigms. In this paper, we presented a graph-based model for describing the relationships of microblog users, in which the distributed storage and map/reduce programs were incorporated. Then, we proposed the map/reduce based algorithm for hierarchical categorization of microblog users based on the concepts of chordal sub graph and join tree in the graph theory. Thus, the categories of microblog users with overlapping and hierarchical properties in various abstraction hierarchies can be obtained flexibly. Experimental results show the feasibility of our method.