Research On Static Overlapping Community Delineation Algorithm Based On Edge Adjacency Relationship

The study of the objective world is a large and complex project,and researchers have been searching for a more efficient and effective research method in order to gain a deeper understanding of the essence of the objective world.Among them,delineating the structure of community networks is a widely adopted and powerful tool.The delineation and study of communities not only helps to understand the nature of the system,but also helps to discover the underlying patterns and laws.Since overlapping communities are more relevant to real life and can more accurately describe the structure of biological and social networks,more and more researchers have shifted their research focus to the discovery and clustering of overlapping communities.However,the existing algorithms for overlapping communities have various problems,such as poor segmentation,poor classification of isolated points,and extensive manual intervention in parameter selection.In order to optimize and solve the above problems,two innovative community discovery algorithms are proposed in this paper for further improvement.The specific work of this paper is as follows:(1)Considering the connectivity of each edge and the network structure between neighbors,an innovative community discovery similarity calculation method is proposed based on the LHN similarity algorithm is chosen,while an innovative community discovery similarity calculation method is paired with the use of RA metrics in order to improve its limitations that may ignore node degree information and to more comprehensively consider the location and attribution information of nodes in the network as well as the path information between nodes,and then based on this algorithm,an edge clustering algorithm based on edge adjacency relationship is proposed.By comparing with the standard division,it can be found that the accuracy of this algorithm reaches 92.4%,while its EQ and NMI indexes have been improved compared with the comparison algorithm.(2)Based on the above-mentioned idea of edge neighborhood relationship,considering that this similarity algorithm,although it makes good use of the properties of edge communities,also has its limitation of not being able to make full use of all topological information between nodes,and also has some inaccuracy for the effect of dividing isolated points,therefore,this paper tries to combine the Jaccard similarity calculation method and NMD index by comparing the same neighborhood between nodes The number of nodes,the number of different neighboring nodes and the average distance of the shortest path between nodes are compared to measure the similarity between nodes more comprehensively,and an innovative community discovery similarity calculation method is proposed,and then a global-based sparse community discovery algorithm is proposed based on this.By comparing with the standard division it can be found that the accuracy of this algorithm reaches 95.9%,and also its EQ and NMI metrics have been improved compared with the comparison algorithm.