Incremental dynamic community discovery algorithm based on multi-core nodes

Abstract

A new incremental dynamic community discovery algorithm MCNIDCD based on multiple core nodes was proposed to address challenges in dynamic community discovery.It adapted to sudden events like the emergence or disappearance of communities during evolution.MCNIDCD categorized core nodes into diffusion and cohesion types, and devised four incremental updating strategies.It adjusted node community membership locally and optimized community structure using an incremental modularity method to facilitate community merging.Evaluation on artificial and real networks shows MCNIDCD’s high conformity to community evolution patterns.In real network experiments, MCNIDCD exhibits a 28% average improvement in modularity performance and significant stability advantages.Its superiority is important for studying dynamic community evolution.

Keywords

• core node;incremental;network evolution;community discovery