Overlapping community detection using weighted consensus clustering

Many overlapping community detection algorithms have been proposed. Most of them are unstable and behave non-deterministically. In this paper, we use weighted consensus clustering for combining multiple base covers obtained by classic non-deterministic algorithms to improve the quality of the result...

Full description

Saved in:
Bibliographic Details
Published in:Pramāṇa 2016-10, Vol.87 (4), p.1-6, Article 58
Main Authors: YANG, LINTAO, YU, ZETAI, QIAN, JING, LIU, SHOUYIN
Format: Article
Language:English
Subjects:
Algorithms
Astronomy
Astrophysics and Astroparticles
Clustering
Communities
Network reliability
Observations and Techniques
Physics
Physics and Astronomy
Reliability analysis
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:Many overlapping community detection algorithms have been proposed. Most of them are unstable and behave non-deterministically. In this paper, we use weighted consensus clustering for combining multiple base covers obtained by classic non-deterministic algorithms to improve the quality of the results. We first evaluate a reliability measure for each community in all base covers and assign a proportional weight to each one. Then we redefine the consensus matrix that takes into account not only the common membership of nodes, but also the reliability of the communities. Experimental results on both artificial and real-world networks show that our algorithm can find overlapping communities accurately.
ISSN:0304-4289
0973-7111
DOI:10.1007/s12043-016-1270-2