Enhanced Community Detection Using Label Propagation Algorithm Integrated with Particle Swarm Optimization

• Author(s): D. Dhanalakshmi ; G. Rajendran
• Paper ID: 1706423
• Page: 412-426
• Published Date: 24-10-2024
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 8 Issue 4 October-2024

Abstract

Community detection in complex networks is pivotal for understanding the structural and functional properties of various systems ranging from social networks to biological systems. Traditional algorithms like the Label Propagation Algorithm (LPA) offer computational efficiency but often suffer from instability and accuracy issues. To address these challenges, this paper introduces the Enhanced Community Detection Using Label Propagation Algorithm with Particle Swarm Optimization (ECDLPA-PSO). By integrating the explorative capabilities of Particle Swarm Optimization (PSO) with LPA, the proposed method aims to enhance the stability and accuracy of community detection. Comparative analyses were conducted against established algorithms, including Girvan-Newman, K-Cliques, Chinese Whispers, Enhanced Community Detection Using Label Propagation Algorithm with ACO (ECDLPA-ACO), and Enhanced Community Detection Using Louvain Algorithm with ACO (ECDLA-ACO). Evaluations based on Modularity, Normalized Mutual Information (NMI), and Execution Time was performed on diverse datasets such as Reddit Hyperlink Network (RH-NW), Amazon Co-purchasing Network (ACP-NW), DBLP Collaboration Network (DBLP-NW), and Twitch Gamers Network (TG-NW). The results demonstrate that ECDLPA-PSO consistently outperforms its counterparts, achieving higher modularity and NMI scores while maintaining competitive execution times. This study underscores the potential of hybrid approaches in advancing community detection methodologies.

Keywords

Community Detection, Particle Swarm Optimization (PSO), ECDLPA-PSO, Modularity, Normalized Mutual Information (NMI), Complex Networks, Optimization Algorithms, Social Network Analysis

Citations

IRE Journals:
D. Dhanalakshmi , G. Rajendran "Enhanced Community Detection Using Label Propagation Algorithm Integrated with Particle Swarm Optimization" Iconic Research And Engineering Journals Volume 8 Issue 4 2024 Page 412-426

IEEE:
D. Dhanalakshmi , G. Rajendran "Enhanced Community Detection Using Label Propagation Algorithm Integrated with Particle Swarm Optimization" Iconic Research And Engineering Journals, 8(4)