Overlapping Community Detection in Complex Networks Using Multi-Objective Particle Swarm Optimization Algorithm
Dhuha Abdulhadi Abduljabbar · 2024
Understanding the structures and functions of networks requires the ability to identify the hidden communities within them. In most real-world networks, research has revealed that nodes can have membership in more than one community, demonstrating the existence of overlapped community structures. As a result, overlap is a significant characteristic of networks, and overlapping community detection has attracted increasing attention recently. In this article, we present a multi-objective optimization algorithm to detect the overlapping topological structure of community in complex networks using the recent framework of particle swarm optimization algorithm. We have proposed two metrics to evaluate overlapping community structure that exhibit a negative correlation. These metrics are known as negative partition density (NPD) and community unfitness (CUF), to serve as optimization objectives. Additionally, our algorithm adopts a line graph representation approach, enabling the generation of overlapping graph partitions from the original interaction graph, thus allowing nodes to belong to multiple clusters. Finally, the proposed algorithm has been tested on both synthetic and real- world social networks, and its results have been compared against well-known state of the art algorithms for overlapping community detection problem. The experimental results demonstrate the effectiveness and efficiency of our proposed algorithm.