Clustering Through Hybrid Network Architecture With Support Vectors

Emrah Ergul, Nafiz Arica, Narendra Ahuja, Sarp Erturk

Research output: Contribution to journalArticle

Abstract

In this paper, we propose a clustering algorithm based on a two-phased neural network architecture. We combine the strength of an autoencoderlike network for unsupervised representation learning with the discriminative power of a support vector machine (SVM) network for fine-tuning the initial clusters. The first network is referred as prototype encoding network, where the data reconstruction error is minimized in an unsupervised manner. The second phase, i.e., SVM network, endeavors to maximize the margin between cluster boundaries in a supervised way making use of the first output. Both the networks update the cluster centroids successively by establishing a topology preserving scheme like self-organizing map on the latent space of each network. Cluster fine-tuning is accomplished in a network structure by the alternate usage of the encoding part of both the networks. In the experiments, challenging data sets from two popular repositories with different patterns, dimensionality, and the number of clusters are used. The proposed hybrid architecture achieves comparatively better results both visually and analytically than the previous neural network-based approaches available in the literature.

Original languageEnglish (US)
Article number7442845
Pages (from-to)1373-1385
Number of pages13
JournalIEEE Transactions on Neural Networks and Learning Systems
Volume28
Issue number6
DOIs
StatePublished - Jun 2017

Fingerprint

Network architecture
Support vector machines
Tuning
Neural networks
Self organizing maps
Clustering algorithms
Topology
Experiments

Keywords

  • Autoencoder (AE) network
  • clustering neural networks
  • greedy layerwise learning
  • prototype encoding (PE) network
  • support vector machine (SVM)

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this

Clustering Through Hybrid Network Architecture With Support Vectors. / Ergul, Emrah; Arica, Nafiz; Ahuja, Narendra; Erturk, Sarp.

In: IEEE Transactions on Neural Networks and Learning Systems, Vol. 28, No. 6, 7442845, 06.2017, p. 1373-1385.

Research output: Contribution to journalArticle

@article{ae7ed73af2484bb09a3d057567758821,
title = "Clustering Through Hybrid Network Architecture With Support Vectors",
abstract = "In this paper, we propose a clustering algorithm based on a two-phased neural network architecture. We combine the strength of an autoencoderlike network for unsupervised representation learning with the discriminative power of a support vector machine (SVM) network for fine-tuning the initial clusters. The first network is referred as prototype encoding network, where the data reconstruction error is minimized in an unsupervised manner. The second phase, i.e., SVM network, endeavors to maximize the margin between cluster boundaries in a supervised way making use of the first output. Both the networks update the cluster centroids successively by establishing a topology preserving scheme like self-organizing map on the latent space of each network. Cluster fine-tuning is accomplished in a network structure by the alternate usage of the encoding part of both the networks. In the experiments, challenging data sets from two popular repositories with different patterns, dimensionality, and the number of clusters are used. The proposed hybrid architecture achieves comparatively better results both visually and analytically than the previous neural network-based approaches available in the literature.",
keywords = "Autoencoder (AE) network, clustering neural networks, greedy layerwise learning, prototype encoding (PE) network, support vector machine (SVM)",
author = "Emrah Ergul and Nafiz Arica and Narendra Ahuja and Sarp Erturk",
year = "2017",
month = "6",
doi = "10.1109/TNNLS.2016.2542059",
language = "English (US)",
volume = "28",
pages = "1373--1385",
journal = "IEEE Transactions on Neural Networks and Learning Systems",
issn = "2162-237X",
publisher = "IEEE Computational Intelligence Society",
number = "6",

}

TY - JOUR

T1 - Clustering Through Hybrid Network Architecture With Support Vectors

AU - Ergul, Emrah

AU - Arica, Nafiz

AU - Ahuja, Narendra

AU - Erturk, Sarp

PY - 2017/6

Y1 - 2017/6

N2 - In this paper, we propose a clustering algorithm based on a two-phased neural network architecture. We combine the strength of an autoencoderlike network for unsupervised representation learning with the discriminative power of a support vector machine (SVM) network for fine-tuning the initial clusters. The first network is referred as prototype encoding network, where the data reconstruction error is minimized in an unsupervised manner. The second phase, i.e., SVM network, endeavors to maximize the margin between cluster boundaries in a supervised way making use of the first output. Both the networks update the cluster centroids successively by establishing a topology preserving scheme like self-organizing map on the latent space of each network. Cluster fine-tuning is accomplished in a network structure by the alternate usage of the encoding part of both the networks. In the experiments, challenging data sets from two popular repositories with different patterns, dimensionality, and the number of clusters are used. The proposed hybrid architecture achieves comparatively better results both visually and analytically than the previous neural network-based approaches available in the literature.

AB - In this paper, we propose a clustering algorithm based on a two-phased neural network architecture. We combine the strength of an autoencoderlike network for unsupervised representation learning with the discriminative power of a support vector machine (SVM) network for fine-tuning the initial clusters. The first network is referred as prototype encoding network, where the data reconstruction error is minimized in an unsupervised manner. The second phase, i.e., SVM network, endeavors to maximize the margin between cluster boundaries in a supervised way making use of the first output. Both the networks update the cluster centroids successively by establishing a topology preserving scheme like self-organizing map on the latent space of each network. Cluster fine-tuning is accomplished in a network structure by the alternate usage of the encoding part of both the networks. In the experiments, challenging data sets from two popular repositories with different patterns, dimensionality, and the number of clusters are used. The proposed hybrid architecture achieves comparatively better results both visually and analytically than the previous neural network-based approaches available in the literature.

KW - Autoencoder (AE) network

KW - clustering neural networks

KW - greedy layerwise learning

KW - prototype encoding (PE) network

KW - support vector machine (SVM)

UR - http://www.scopus.com/inward/record.url?scp=85028316423&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028316423&partnerID=8YFLogxK

U2 - 10.1109/TNNLS.2016.2542059

DO - 10.1109/TNNLS.2016.2542059

M3 - Article

C2 - 28113825

AN - SCOPUS:85028316423

VL - 28

SP - 1373

EP - 1385

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

SN - 2162-237X

IS - 6

M1 - 7442845

ER -