Research Of Community Detection’s Algorithm Based On Friends Similarity From Online Social Networks

“Birds of a feather flock together”. So the community structure is one of the mostcommon and most important topological properties in online social networks. Thecommunity could be loosely described as a sub-graph which has a high density of edgeswithin them, and a lower density of edges between groups in online social networks. In thesame community, they may share the same tastes and interests. It is very important todiscover the potential communities with their hierarchical structures in online socialnetworks accurately and effectively. Advancing of precision of community structure ofonline social network is important in research and practice.Firstly, this paper discusses the online social network research background, the studyof current conditions, and the main research contents. Then, the paper focus on somerepresentative algorithm about online social networks recognition, such as GN algorithm,CNM algorithm, Louvain algorithm etc. To partition online social networks into groupsfast and correctly, after analyzing the existing methods, this paper proposes two algorithmsfor discovering community structures in online social networks based on the principle oftriadic closure, as show below.An algorithm for mining local community based on node s neighbors similarity fromcomplex networks. This paper proposes a new method to discover the local communitystructure. To start with the maximal-degree node and get the initial local community by thedefinition of shared neighbors. Then leverage the maximal similarity of node and the localcommunity s Q value and update the initial local community. The experiments on threetypical complex networks show that the algorithm can effectively mine the intrinsic localcommunity structure in networks.A new community structure detection method based on central nodes similarity. Thebasic idea is, with the central nodes and friends similarity as criteria, based on a set ofcentral nodes and using the node’s local information efficiently to find the communitystructure of online social networks. The proposed algorithm is based on the local structure,so it’s very efficient. NMI is used to evaluate the result of the proposed algorithm, which is also compared with LPA and CNM algorithm. NMI is a criterion widely used forcommunity hierarchy detection methods. When being tested on some typical real-worldand computer-generated networks, this algorithm demonstrates excellent detection resultsand very fast processing performance, much better than the existing comparablealgorithms of the same kind.