Community Detection Based On Autonomy-Oriented Computing And Analysis Of Community Vulnerability

Community refers to a set of nodes with close connections or similar attributes in a complex network.Usually,the edges between communities are sparse,while the edges within communities are closely connected.Community is an important structural attribute of the network system.Community detection is helpful to mining the key information in the network and analyzing the network structure characteristics.Therefore,the research of community detection method has important practical significance.At the same time,the security of the community in the network has become particularly important.The vulnerability assessment of the community can find the vulnerable links in the network in time,and help to develop effective measures to prevent the network from suffering large-scale losses.Therefore,it is indispensable to study the vulnerability analysis of the community.Based on the above problems,this paper proposes a community detection method based on autonomic computing and a new community vulnerability analysis method,and takes the real world network as the research object,and verifies the effectiveness of the method through a large number of simulation experiments.The main work contents and innovative achievements of this paper are as follows:(1)Research on community detection method based on autonomic computingThis paper presents a new community detection method(AOCCD)based on autonomic computing.Based on the Autonomy-oriented computing(AOC)system,the method mainly includes two stages: detection and merging.In the detection stage,the autonomous bodies in the system excavate the local community of their affiliated nodes respectively,and determine which node can join the local community through the local similarity index and the local module degree gain index.In the merge stage,all autonomous bodies have completed their community detection tasks,and the nodes with the same community label can be merged to complete the preliminary division of communities.Finally,merge some "small clubs" using the merge strategy presented in this article.Experiments show that AOCCD algorithm is more accurate in the division of network community,the community quality is higher,and better than some traditional algorithms.(2)An analytical study of community vulnerability considering the internal and external characteristics of communityConsidering the internal and external characteristics of community,this paper proposes a new method of community vulnerability analysis.The internal characteristics of the community include the average degree of cut points,the connection density and the weight between edges.The external features include the number of edges and the weight of edges between communities.Based on the above five characteristics,this paper defines the community vulnerability index and quantifies the community vulnerability.The higher the community vulnerability value,the more vulnerable the community will be.The feasibility of the community vulnerability assessment method proposed in this paper is verified by experiments in real world network.The method proposed in this paper can accurately assess the vulnerability of the community and timely discover the vulnerable links in the network,so as to prevent the network from suffering large-scale losses.