Personalized Community Discovery Algorithm Based on GN

Abstract

Nowadays, complex network has become a hot topic with the development of Internet. How to further explore the structure of complex networks also becomes a current focus. Generally, network has a strong community structure, so we could find these community structures to help us understand characteristics of complex networks. As a representative of classic community discovery algorithms, GN algorithm could discover complex network communities. Traditional GN algorithm removes the edge with highest betweenness in networks, which gets partion results. This paper improves traditional GN algorithm by quantifying the relationship among users based on micro-blog. Our improved GN algorithm that removes edges with the lowest relationship applies to asymmetric networks. Verified by real social networks and simulated micro-blog networks, the improved GN algorithm could discover community structure more precisely and effectively. Compared with traditional GN algorithm, running time of our algorithm is less, which reduces time complexity.