Identifying overlapping communities in social networks using multi-scale local information expansion

TL;DR

This paper introduces a novel, efficient algorithm for detecting overlapping communities in social networks by optimizing topological entropy and modeling multi-scale social interactions without requiring complete graph information.

Contribution

The proposed method is the first to optimize topological entropy for community detection, supporting overlapping communities and multi-scale analysis with low computational complexity.

Findings

Supports overlapping community detection.

Supports multi-scale community analysis.

High efficiency and accuracy in real networks.

Abstract

Most existing approaches for community detection require complete information of the graph in a specific scale, which is impractical for many social networks. We propose a novel algorithm that does not embrace the universal approach but instead of trying to focus on local social ties and modeling multi-scales of social interactions occurring in those networks. Our method for the first time optimizes the topological entropy of a network and uncovers communities through a novel dynamic system converging to a local minimum by simply updating the membership vector with very low computational complexity. It naturally supports overlapping communities through associating each node with a membership vector which describes node's involvement in each community. This way, in addition to uncover overlapping communities, we can also describe different multi-scale partitions by tuning the characteristic size of modules from the optimal partition. Because of the high efficiency and accuracy of the algorithm, it is feasible to be used for the accurate detection of community structure in real networks.