Network community discovery method based on optimal network structure

Abstract

The invention discloses a network community discovery method based on an optimal network structure, and the method comprises the steps: 1), generating the optimal network structure for the discovery of a network community from an original network structure; 2) based on the optimal network structure, learning vector representation for network nodes; and 3) based on the network node representation,dividing the network nodes by using a machine learning clustering algorithm so as to obtain a community structure of the network. Compared with the existing community discovery method, the network community discovery method based on the optimal network structure disclosed by the invention can be self-adaptive to various network types in the real world through a data-driven strategy, and has higherapplication universality; meanwhile, the adverse effect of inter-community edge connection of a fuzzy community structure in an original network structure on community discovery is considered and eliminated, more accurate and more stable community discovery is performed through an optimal network structure with a finer-grained edge connection relationship, and the method has the advantage of practical application.

Description

technical field [0001] The invention belongs to the field of complex network analysis, and in particular relates to a network community discovery method based on an optimal network structure. Background technique [0002] Community structure is a common structural phenomenon in complex networks, and it is ubiquitous in social networks, biological networks, and academic paper citation networks. Discovering and identifying the community structure in the network is the basis for in-depth analysis of network data. It has two main functions: first, it can analyze the macro-connection and interaction relationships emerging between nodes in the network from a global perspective, rather than limited Secondly, in large-scale network analysis, the network is divided into multiple sub-networks for analysis and aggregation through community discovery, which can eliminate the difficulties caused by the large-scale analysis of the entire network. [0003] The community discovery task on ...

AI Technical Summary

Problems solved by technology

[0005] The above-mentioned network community discovery methods are all based on the original binary network structure, without considering the noise edges of the fuzzy community structure in the original network structure, and without mining the finer-grained edge connection relationship in the network

At the same time, the above methods are traditional community discovery methods, all of which use fixed network partition metrics, and cannot be widely applied to real-world networks with different structures.

Images