Loading…

Novel heuristic density-based method for community detection in networks

Recent years have witnessed a growing recognition on the community detection in networks. Diverse techniques have been devoted to uncovering community structures in complex networks and amongst which are the density-based methods. Density-based avenues are very popular in data clustering field. They...

Full description

Saved in:
Bibliographic Details
Published in:Physica A 2014-06, Vol.403, p.71-84
Main Authors: Gong, Maoguo, Liu, Jie, Ma, Lijia, Cai, Qing, Jiao, Licheng
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Recent years have witnessed a growing recognition on the community detection in networks. Diverse techniques have been devoted to uncovering community structures in complex networks and amongst which are the density-based methods. Density-based avenues are very popular in data clustering field. They rely on two parameters which are utilized by us to process the community detection problem. In this paper, a novel view to look deep into the network structure from the community level is tested and a heuristic density-based approach for community detection is put forward. In the proposed method, firstly, both of the two parameters are under consideration and all the possible parameter pairs are exploited. These parameter pairs produce all kinds of partitions through the classic method. Secondly, these partitions are processed by our proposed strategy consisting of classification, mergence, decomposition and recombination. After employing the proposed strategy, a community division with high quality is uncovered. Experiments on both synthetic and real-world networks demonstrate the effectiveness of the proposed method. •We adopt the density-based clustering method to reveal community structure in networks.•The drawback about the parameters of the density-based method has been alleviated.•The parameters are utilized to analyze the network from multiple degrees.•The proposed method provides a novel platform to generate network structures on the community level.
ISSN:0378-4371
1873-2119
DOI:10.1016/j.physa.2014.01.043