Label Ring Propagation Overlapping Community Identification Method Based On Node Membership Degree

With the rapid development of Internet technology and the widespread popularization of mobile smart terminals,social networks composed of typical social applications(such as Weibo,We Chat,Facebook,etc.)have gradually become important media and main means for people to communicate,communicate and express emotions.Social networks are constantly changing human behavior patterns and social forms,affecting people's lives and social and economic development.Virtual community identification takes the analysis of social network structure characteristics as the main research target,and is the basis for the study of social network characteristics and functions.Among them,the overlapping community identification method has become a research hotspot in the current academic community because it can mine the overlapping structure of the community and more in line with the actual application needs.Research on it is helpful for people to understand society and explain social phenomena,and to reveal the internal structural functions of social networks.It has important theoretical research value and broad application prospects.In the context of the rapid increase in the amount of data in the current Internet environment,traditional overlapping community recognition algorithms are not well suited for overlapping community recognition scenarios,and the quality of community recognition is poor and the stability is weak.A node-based membership is proposed.Degree of Labeled Circular Propagation Overlap Community Identification Algorithm.The algorithm first makes full use of the ring structure in the social network for label propagation to discover the community structure in the network.The node's influence is measured by the topological potential of the nodes,and the nodes are selected according to the node's influence in the node selection phase of the algorithm to effectively eliminate The randomness of the algorithm is defined,and the node membership is defined according to the node topological potential and the local similarity between nodes.The label truncation strategy is implemented through the concept of node membership to avoid the problem of artificial overlap.Finally,experiments are performed on real and artificial data sets.Experimental results show that the algorithm proposed in this paper has high effectiveness and stability,and can effectively improve the stability and quality of community recognition results without the need to manually set parameters.