A Parallel and Scalable Framework for Non-overlapping Community Detection Algorithms

Songchang Jin, Yuchao Zhang, Yuanping Nie, Xiang Zhu, Hong Yin, Aiping Li, Shuqiang Yang

Abstract: Community detection has been wildly studied during the past years by varies of researchers, and a plenty of excellent algorithms and approaches have been proposed. But networks are becoming larger and higher complicated in nowadays. How to excavate the hidden community structures in the expanding networks quickly with existing excellent methods has become a challenge. In this paper, we designed a parallel community discovery framework based on Map-reduce and implemented parallel version of some excellent existing standalone community detection methods. Results of empirical tests show that the framework is able to significantly speed up the mining process without compromising the accuracy excessively.