A detection of overlapping community in mobile social network

Detecting a community structure reduces a complexity of networks consisted of large datasets, which is important to manage and understand core groups of the network. Especially, to accurately find a community structure, it is necessary to analyze and consider characteristics of a target network. Mobile social network is a weighted and directed graph. All links in the network have different weights and directions according to social interaction. Also a node should belong to several communities simultaneously. Conventional methods for detecting community structures in social network ignore these properties. Therefore, in this paper, we propose a method for detecting overlapping communities in mobile social network. This method is composed of two stages. First stage is to find community structures with directed and weighted link clustering based on standard single-linkage hierarchical clustering. Second stage is to evaluate the community structures to detect optimal mobile communities. Then we show that communities detected our method satisfies the features of real social communities such as node overlapping and hierarchical structure.