Kernel spectral clustering for big data networks

RaghvenPhDa Mall, Rocco Langone, Johan A.K. Suykens

Research output: Contribution to journalArticle

52 Citations (Scopus)

Abstract

This paper shows the feasibility of utilizing the Kernel Spectral Clustering (KSC) method for the purpose of community detection in big data networks. KSC employs aprimal-dual framework to construct a model. It results in a powerful property of effectively inferring the community affiliation for out-of-sample extensions. The original large kernelmatrix cannot fitinto memory. Therefore, we select a smaller subgraph that preserves the overall community structure to construct the model. It makes use of the out-of-sampleextension property for community membership of the unseen nodes. We provide anovel memory- and computationally efficient model selection procedure based on angular similarity in the eigenspace. We demonstrate the effectiveness of KSC on large scalesynthetic networks and real world networks like the YouTube network, a road network ofCalifornia and the Livejournal network. These networks contain millions of nodes and several million edges.

Original languageEnglish
Pages (from-to)1567-1586
Number of pages20
JournalEntropy
Volume15
Issue number5
DOIs
Publication statusPublished - 1 May 2013
Externally publishedYes

Fingerprint

spectral methods
roads

Keywords

  • Angular similarity
  • Kernel spectral clustering
  • Out-of-sample extensions
  • Sampling graphs

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Mall, R., Langone, R., & Suykens, J. A. K. (2013). Kernel spectral clustering for big data networks. Entropy, 15(5), 1567-1586. https://doi.org/10.3390/e15051567

Kernel spectral clustering for big data networks. / Mall, RaghvenPhDa; Langone, Rocco; Suykens, Johan A.K.

In: Entropy, Vol. 15, No. 5, 01.05.2013, p. 1567-1586.

Research output: Contribution to journalArticle

Mall, R, Langone, R & Suykens, JAK 2013, 'Kernel spectral clustering for big data networks', Entropy, vol. 15, no. 5, pp. 1567-1586. https://doi.org/10.3390/e15051567
Mall, RaghvenPhDa ; Langone, Rocco ; Suykens, Johan A.K. / Kernel spectral clustering for big data networks. In: Entropy. 2013 ; Vol. 15, No. 5. pp. 1567-1586.
@article{c239ce1958a346e1a3d6bc6131554ce4,
title = "Kernel spectral clustering for big data networks",
abstract = "This paper shows the feasibility of utilizing the Kernel Spectral Clustering (KSC) method for the purpose of community detection in big data networks. KSC employs aprimal-dual framework to construct a model. It results in a powerful property of effectively inferring the community affiliation for out-of-sample extensions. The original large kernelmatrix cannot fitinto memory. Therefore, we select a smaller subgraph that preserves the overall community structure to construct the model. It makes use of the out-of-sampleextension property for community membership of the unseen nodes. We provide anovel memory- and computationally efficient model selection procedure based on angular similarity in the eigenspace. We demonstrate the effectiveness of KSC on large scalesynthetic networks and real world networks like the YouTube network, a road network ofCalifornia and the Livejournal network. These networks contain millions of nodes and several million edges.",
keywords = "Angular similarity, Kernel spectral clustering, Out-of-sample extensions, Sampling graphs",
author = "RaghvenPhDa Mall and Rocco Langone and Suykens, {Johan A.K.}",
year = "2013",
month = "5",
day = "1",
doi = "10.3390/e15051567",
language = "English",
volume = "15",
pages = "1567--1586",
journal = "Entropy",
issn = "1099-4300",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "5",

}

TY - JOUR

T1 - Kernel spectral clustering for big data networks

AU - Mall, RaghvenPhDa

AU - Langone, Rocco

AU - Suykens, Johan A.K.

PY - 2013/5/1

Y1 - 2013/5/1

N2 - This paper shows the feasibility of utilizing the Kernel Spectral Clustering (KSC) method for the purpose of community detection in big data networks. KSC employs aprimal-dual framework to construct a model. It results in a powerful property of effectively inferring the community affiliation for out-of-sample extensions. The original large kernelmatrix cannot fitinto memory. Therefore, we select a smaller subgraph that preserves the overall community structure to construct the model. It makes use of the out-of-sampleextension property for community membership of the unseen nodes. We provide anovel memory- and computationally efficient model selection procedure based on angular similarity in the eigenspace. We demonstrate the effectiveness of KSC on large scalesynthetic networks and real world networks like the YouTube network, a road network ofCalifornia and the Livejournal network. These networks contain millions of nodes and several million edges.

AB - This paper shows the feasibility of utilizing the Kernel Spectral Clustering (KSC) method for the purpose of community detection in big data networks. KSC employs aprimal-dual framework to construct a model. It results in a powerful property of effectively inferring the community affiliation for out-of-sample extensions. The original large kernelmatrix cannot fitinto memory. Therefore, we select a smaller subgraph that preserves the overall community structure to construct the model. It makes use of the out-of-sampleextension property for community membership of the unseen nodes. We provide anovel memory- and computationally efficient model selection procedure based on angular similarity in the eigenspace. We demonstrate the effectiveness of KSC on large scalesynthetic networks and real world networks like the YouTube network, a road network ofCalifornia and the Livejournal network. These networks contain millions of nodes and several million edges.

KW - Angular similarity

KW - Kernel spectral clustering

KW - Out-of-sample extensions

KW - Sampling graphs

UR - http://www.scopus.com/inward/record.url?scp=84878361008&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84878361008&partnerID=8YFLogxK

U2 - 10.3390/e15051567

DO - 10.3390/e15051567

M3 - Article

AN - SCOPUS:84878361008

VL - 15

SP - 1567

EP - 1586

JO - Entropy

JF - Entropy

SN - 1099-4300

IS - 5

ER -