Overlapping Community Discovery Algorithm Based on Local Extension

Abstract

Detecting communities in social networks represents a significant task in understanding the structures and functions of networks. However, in real graphs vertices are often shared between communities, thus overlapping mining in the network community is more meaningful. In this paper, we propose an effective method for overlapping community discovery, which is to discover the initial core community and extend the core community until it reaches a stable condition. Through experiments on several real datasets, this method can effectively partition the overlapping communities.