Density maximization overlapping community discovery method and system based on local structure

Abstract

The invention provides a density maximization overlapping community discovery method based on a local structure. The density maximization overlapping community discovery method comprises: a step A, abstracting a community to be divided into an abstract network formed by connecting nodes; a step B, obtaining the degree of each node, and constructing an initialized community by using the node with the maximum degree and neighbor nodes of the node with the maximum degree; a step C, repeating the step B for the remaining nodes until no undivided nodes exist in the abstract network; a step D, sequentially judging whether the nodes in the small communities can be added into the large communities or not, and expanding the initialized communities; and E, judging whether the expanded communities meet a merging condition or not, merging the communities meeting the merging condition, and outputting a community division result. The invention further provides a system based on the method. The method has the advantages that the initialized communities are determined through selection of the seed nodes, whether the communities can be combined or not is judged according to the community expansionresult, and the accuracy and the expansion efficiency of the community division result are improved.

Description

technical field [0001] The present invention relates to the technical field of community division, in particular to a method and system for discovering overlapping communities with maximum density based on local structure. Background technique [0002] Networks exist widely in our real life, such as social networks in social systems, protein interaction networks in biological systems, the World Wide Web and the Internet in computer networks, etc. Each individual in these networks communicates with other individuals through information exchange channels connect. With the in-depth research on the network, it is found that there are community structures in many actual networks, that is, the entire network is composed of several communities. For example, a network that exists in reality, such as each device in a computer network, can be abstracted into a node, and the network channels between nodes can be abstracted into edges between nodes. It can be found that the nodes in ea...

AI Technical Summary

Problems solved by technology

However, the membership degree relationship considered in the expansion process of this application only considers the degree of nodes in the local community and the relationship between the degree of the entire abstract network, and does not consider the connection relationship between nodes in the local community structure and the influence of nodes on the entire community structure. , unable to solve the problem of large-scale network sparseness

Images