Partitioning Algorithm Based On The Social Networking Community

Social network can be used to describe the social actual network, which includes the social relations between people, prey relations between species, cooperation relations in science. Numerous studies have shown that different social networks have many common structural features in the real world. For example,small world property, no scaling, community structure. the research of community structure is widely concerned by all subject researchers. They found that the same knowledge points, the concept of the same subject are linked more closely, and different areas knowledge are linked relatively sparse. this paper focus on the study of the community structure,whose nature can help us understand the network structure and analyse network characteristics. It is great significance for us to recognize and understand true structure and function of complex networks by detectiding community structure. Currently, many researchers have been engaged to the detection of community structure and proposed variety of algorithms to quickly and accurately find the community structure in the network, some classical algorithms,such as the GN algorithm, Kernighan-Lin algorithm. But contradiction between the accuracy and the algorithm time complexity of the division of community structure is still a major problem. In this paper,we analysis the community structure algorithm and compare the advantages and disadvantages of the algorithms.on this basis, we propose two improveed algorithms .This article make the following work:(1) Studying the basic knowledge of social network theory, introducing some representations and some basic concepts of social network, Then introduce several classic algorithms, and summarize their advantages and disadvantages.(2) A new algorithm is constructed by the distance difference between adjacent nodes--the algorithm based on division of the adjacent nodes clustering. First ,find the initial cluster centers for each community, the M nodes as the initial community, calculate the smallest distance with the first node ,put it into the appropriate community. The algorithm is applied to network instance ,the algorithm has a high accuracy.(3) Determine the similarity between nodes by calculating the similarity coefficient.The Tierarchical clustering is based on similarity of the nodes in the same community. The nodes are grouped according to the similarity and bottom-up strategy,First, each node as a community, then merge these initial community into the growing community .Finally stop the algorithm by modularity Q.Innovation of this paper:(1) The algorithm based on division of the adjacent nodes clustering define the distance difference between adjacent nodes, this parameter is expressed quantitatively degree of difference between different community nodes.(2) Similarity of two nodes in the network, hierarchical clustering algorithm and modularity construct an algorithm, this algorithm has a higher accuracy.Partitioning algorithm of the paper proposed have many imperfections,related work remains to be further studied.