Social network sampling with keeping community structure

摘要

Online social networks (OSNs) have become an important way to connect people in recent years. This phenomenon generates lots of interest in graph analysis. Because of the large scale of OSNs, it is almost impossible to perform the research with complete datasets. For some research, i.e. the transition of message, the community structure and shortest path are both important issues. In this paper, we propose a sampling method which can keep the properties of network community structure and shortest paths.