Detecting The Overlapping Community Structure In Social Networks

The development of Internet technology promote the progress of online social network applications, and the popular social networking herald the arrival of the era of big data. The study found that online social networking is a microcosm of the real social network; they all have a community structure characteristic. Analysis, do research and in-depth study the structural characteristics and the evolution of them can get insights into the complex social network system then guide the practical problem. Community discovery is a technology of community divided in the research of complex network and it including allow and does not allow overlapping point. The study found that overlapping communities was more in line with the real social network; a node may belong to multiple communities. The study found that online social network also has this feature. By the thinking of local information and the characteristics of online social networks, This paper presents a center expansion discovery algorithm-CE algorithm based on node influence and contribution factor which can detecting the overlapping point in the communities. This method is constantly absorbing neighbor nodes from the community center nodes (the greatest influence node) to find all members of a community. It does not require pre-designated number of community. In order to be more in line with the characteristics of online social networks, the algorithm introduced the concept of the contribution factor, degree of overlapping and the nodes’bias of community. Paper use karate club Zachary, American College Football and a Sina micro blogging user’s fans network to validate this algorithm. Experiments show that the reliability of the CE algorithm which also can find the points of overlap in the Community.Community discovery algorithms which can find points of overlap are not less, But less or say no which study these overlaps point. Beside detected the overlapping point paper also analysis the overlapping points more biased in favor of belong to which communities. Finally, do a full summary and next job outlook.