Core member identification-based community discovery method

Abstract

The invention discloses a core member identification-based community discovery method. The method comprises the following steps of firstly performing node betweenness center value-based core member identification on nodes in a network, then performing initial clustering, and forming an initial social network structure according to the effects of the nodes in the network; and secondly regarding aninitially clustered node as a node, performing dimension reduction on a similarity matrix M among the nodes to form a new similarity matrix M', and replacing a modularity increment matrix to serve asan input of a CNM algorithm. The core member identification provides a condition for the initial clustering; and the matrix scale can be reduced through the dimension reduction when the similarity among the nodes is calculated, so that the algorithm loss time is shortened. The core member identification algorithm adopts a breadth-first search-based quick betweenness centrality calculation method,and the spatial complexity and the practical complexity of the core member identification algorithm are both lower than those of a conventional algorithm, so that important members in a community network can be quickly located and reasonable division of a social network structure is facilitated.

Description

technical field [0001] The invention relates to a community discovery method based on core member identification. Background technique [0002] A social network refers to a collection of actors and their connections, and a social network usually presents the characteristics of a small group, that is, a community. The analysis results of the virtual social network can be used as a mapping of the real social network. Therefore, community discovery technology, as a key technology of social network analysis, has attracted more and more attention from scholars. However, the current community discovery algorithms still have problems that need to be solved urgently, such as the quality of community discovery is not high or the algorithm operation efficiency is low. [0003] Most of the existing community discovery algorithms are implemented by analyzing the information of each user, extracting feature words from it, and calculating the similarity of other users based on one user....

AI Technical Summary

Problems solved by technology

This type of method is similar to the segmentation method based on the concept of edge intermediary (Edge Btweeness) - GN algorithm, but the GN algorithm has a large time complexity and is not suitable for large-scale networks.

Images