Non-negative sparse autoencoder neural networks for the detection of overlapping, hierarchical communities in networked datasets

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We propose the first use of a non-negative sparse autoencoder (NNSAE) neural network for community structure detection in complex networks. The NNSAE learns a compressed representation of a set of fixed-length, weighted random walks over the network, and communities are detected as subsets of network nodes corresponding to non-negligible elements of the basis vectors of this compression. The NNSAE model is efficient and online. When utilized for community structure detection, it is able to uncover potentially overlapping and hierarchical community structure in large networks.

Original languageEnglish (US)
Article number043141
JournalChaos
Volume22
Issue number4
DOIs
StatePublished - Oct 4 2012

Fingerprint

Community Structure
Complex networks
Overlapping
Non-negative
Neural Networks
Neural networks
Hierarchical Structure
Complex Networks
Random walk
Compression
Subset
random walk
set theory
Vertex of a graph
Community
Model

All Science Journal Classification (ASJC) codes

  • Statistical and Nonlinear Physics
  • Mathematical Physics
  • Physics and Astronomy(all)
  • Applied Mathematics

Cite this

@article{a26f36283470448a956c26fe1f8dd98e,
title = "Non-negative sparse autoencoder neural networks for the detection of overlapping, hierarchical communities in networked datasets",
abstract = "We propose the first use of a non-negative sparse autoencoder (NNSAE) neural network for community structure detection in complex networks. The NNSAE learns a compressed representation of a set of fixed-length, weighted random walks over the network, and communities are detected as subsets of network nodes corresponding to non-negligible elements of the basis vectors of this compression. The NNSAE model is efficient and online. When utilized for community structure detection, it is able to uncover potentially overlapping and hierarchical community structure in large networks.",
author = "Sarah Rajtmajer and Brian Smith and Shashi Phoha",
year = "2012",
month = "10",
day = "4",
doi = "10.1063/1.4771600",
language = "English (US)",
volume = "22",
journal = "Chaos",
issn = "1054-1500",
publisher = "American Institute of Physics Publising LLC",
number = "4",

}

Non-negative sparse autoencoder neural networks for the detection of overlapping, hierarchical communities in networked datasets. / Rajtmajer, Sarah; Smith, Brian; Phoha, Shashi.

In: Chaos, Vol. 22, No. 4, 043141, 04.10.2012.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Non-negative sparse autoencoder neural networks for the detection of overlapping, hierarchical communities in networked datasets

AU - Rajtmajer, Sarah

AU - Smith, Brian

AU - Phoha, Shashi

PY - 2012/10/4

Y1 - 2012/10/4

N2 - We propose the first use of a non-negative sparse autoencoder (NNSAE) neural network for community structure detection in complex networks. The NNSAE learns a compressed representation of a set of fixed-length, weighted random walks over the network, and communities are detected as subsets of network nodes corresponding to non-negligible elements of the basis vectors of this compression. The NNSAE model is efficient and online. When utilized for community structure detection, it is able to uncover potentially overlapping and hierarchical community structure in large networks.

AB - We propose the first use of a non-negative sparse autoencoder (NNSAE) neural network for community structure detection in complex networks. The NNSAE learns a compressed representation of a set of fixed-length, weighted random walks over the network, and communities are detected as subsets of network nodes corresponding to non-negligible elements of the basis vectors of this compression. The NNSAE model is efficient and online. When utilized for community structure detection, it is able to uncover potentially overlapping and hierarchical community structure in large networks.

UR - http://www.scopus.com/inward/record.url?scp=84871890674&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871890674&partnerID=8YFLogxK

U2 - 10.1063/1.4771600

DO - 10.1063/1.4771600

M3 - Article

VL - 22

JO - Chaos

JF - Chaos

SN - 1054-1500

IS - 4

M1 - 043141

ER -