Core Node Knowledge Based Multi-Objective Particle Swarm Optimization for Dynamic Community Detection

Yifei Sun,Xin Sun,Zhuo Liu,Yifei Cao,Jie Yang
DOI: https://doi.org/10.1016/j.cie.2022.108843
IF: 7.18
2023-01-01
Computers & Industrial Engineering
Abstract:Community detection aims to classify nodes with similar attributes, and it becomes more difficult when the network structure changes over time. In order to tap into the evolution of dynamic networks, nodes with high criticality in dynamic networks clearly need to be attended. In this paper, a novel dynamic community detection algorithm based on particle swarm optimization is proposed. The core nodes in the network are obtained by calculating the resistance distance of each node. The nodes associated with these core nodes that make up the constant community. The constant community should be retained is the knowledge gained from the evolution of the core nodes in two consecutive time steps. To ensure the quality of community division at the current time step and the smoothness of two consecutive time steps, a new particle updating parameter is proposed, which is based on the optimal individual at the previous time step and the population at the current time step. Compared with the other well-known related algorithms, experimental result on many real-world and synthetic networks have illustrated the higher accuracy and stability of the proposed.
What problem does this paper attempt to address?