A new selective neural network ensemble with negative correlation

Heesung Lee, Euntai Kim, Witold Pedrycz

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

An ensemble of neural networks exhibits higher generalization performance compared to a single neural network. In this paper, a new design method for a neural network ensemble is proposed. The hierarchical pair competition-based parallel genetic algorithm (HFC-PGA) is employed to train the neural networks forming the ensemble. The aim of the HFC-PGA is to achieve not only the best neural network, but also a diversity of potential neural networks. A set of component neural networks is selected to build an ensemble such that the generalization error is minimized and the negative correlation is maximized. Finally, some experiments are carried out using several data sets to illustrate and quantify the performance of the proposed method.

Original languageEnglish
Pages (from-to)488-498
Number of pages11
JournalApplied Intelligence
Volume37
Issue number4
DOIs
Publication statusPublished - 2012 Dec 1

Fingerprint

Neural networks
Parallel algorithms
Genetic algorithms
Experiments

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence

Cite this

Lee, Heesung ; Kim, Euntai ; Pedrycz, Witold. / A new selective neural network ensemble with negative correlation. In: Applied Intelligence. 2012 ; Vol. 37, No. 4. pp. 488-498.
@article{be751f42d4694b2794fbb2bcfd3df2e9,
title = "A new selective neural network ensemble with negative correlation",
abstract = "An ensemble of neural networks exhibits higher generalization performance compared to a single neural network. In this paper, a new design method for a neural network ensemble is proposed. The hierarchical pair competition-based parallel genetic algorithm (HFC-PGA) is employed to train the neural networks forming the ensemble. The aim of the HFC-PGA is to achieve not only the best neural network, but also a diversity of potential neural networks. A set of component neural networks is selected to build an ensemble such that the generalization error is minimized and the negative correlation is maximized. Finally, some experiments are carried out using several data sets to illustrate and quantify the performance of the proposed method.",
author = "Heesung Lee and Euntai Kim and Witold Pedrycz",
year = "2012",
month = "12",
day = "1",
doi = "10.1007/s10489-012-0342-3",
language = "English",
volume = "37",
pages = "488--498",
journal = "Applied Intelligence",
issn = "0924-669X",
publisher = "Springer Netherlands",
number = "4",

}

A new selective neural network ensemble with negative correlation. / Lee, Heesung; Kim, Euntai; Pedrycz, Witold.

In: Applied Intelligence, Vol. 37, No. 4, 01.12.2012, p. 488-498.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A new selective neural network ensemble with negative correlation

AU - Lee, Heesung

AU - Kim, Euntai

AU - Pedrycz, Witold

PY - 2012/12/1

Y1 - 2012/12/1

N2 - An ensemble of neural networks exhibits higher generalization performance compared to a single neural network. In this paper, a new design method for a neural network ensemble is proposed. The hierarchical pair competition-based parallel genetic algorithm (HFC-PGA) is employed to train the neural networks forming the ensemble. The aim of the HFC-PGA is to achieve not only the best neural network, but also a diversity of potential neural networks. A set of component neural networks is selected to build an ensemble such that the generalization error is minimized and the negative correlation is maximized. Finally, some experiments are carried out using several data sets to illustrate and quantify the performance of the proposed method.

AB - An ensemble of neural networks exhibits higher generalization performance compared to a single neural network. In this paper, a new design method for a neural network ensemble is proposed. The hierarchical pair competition-based parallel genetic algorithm (HFC-PGA) is employed to train the neural networks forming the ensemble. The aim of the HFC-PGA is to achieve not only the best neural network, but also a diversity of potential neural networks. A set of component neural networks is selected to build an ensemble such that the generalization error is minimized and the negative correlation is maximized. Finally, some experiments are carried out using several data sets to illustrate and quantify the performance of the proposed method.

UR - http://www.scopus.com/inward/record.url?scp=84857820925&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857820925&partnerID=8YFLogxK

U2 - 10.1007/s10489-012-0342-3

DO - 10.1007/s10489-012-0342-3

M3 - Article

VL - 37

SP - 488

EP - 498

JO - Applied Intelligence

JF - Applied Intelligence

SN - 0924-669X

IS - 4

ER -