Detecting overlapping communities through the utilization of multi-agent collective intelligence

Andulrahman, Yadgar Sirwan; Hama Ali, Surkew Mohammed

doi:10.24271/psr.2024.450246.1551

Detecting overlapping communities through the utilization of multi-agent collective intelligence

Document Type : Original Article

Authors

yadgar sirwan andulrahman ¹

Surkew Mohammed Hama Ali ²

¹ Information technology, , Kurdistan Technical instituted, Sulaymaniyah, Kurdistan Regional Government, Iraq

² Information technology, Kurdistan Technical instituted , Sulaymaniyah, Kurdistan Regional Government, Iraq

10.24271/psr.2024.450246.1551

Abstract

The perspective of social network analysis provides a clear method for decomposing and analyzing the overall structure of social entities. Detecting communities in networks is a fundamental challenge in the network science, and it is also a major concern after identifying communities to identify the main community members who belong to multiple communities. Finding communities that overlap with each other is an essential and exciting subject in social network recommender systems and data mining. The algorithm presented in this paper utilizes multi-agent particle swarm optimization, demonstrating self-organization among agent activities. The incorporation of collective intelligence enhances the accuracy of the global search. Through the application of a specialized encoding method, the algorithm discerns the count of communities, where modularity index is utilized as the fitness function in particle swarm optimization..

Keywords

Particle Swarm Optimization

Multi-Agent

Modularity

Social Network