Kernel local Fisher discriminant analysis based manifold-regularized SVM model for financial distress predictions

Shian-Chang Huang, Yu-Cheng Tang, Chih Wei Lee, Ming Jen Chang

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Support vector machines (SVM) have demonstrated excellent performance in numerous areas of pattern recognitions. However, traditional SVM does not make efficient use of both labeled training data and unlabeled testing data. Moreover, high dimensional and nonlinear distributed data generally degrade the performance of a classifier due to the curse of dimensionality in financial distress (or bankruptcy) predictions. To address these problems, this study proposes a novel hybrid classifier which integrates Kernel local Fisher discriminant analysis (KLFDA) with a manifold-regularized SVM (MR-SVM). KLFDA is employed to find an optimal projection which maximizes the margin between data points from different classes at each local area of data manifold, while MR-SVM data-dependently warps the structure of feature space to reflect the underlying geometry of the data manifold. Compared with other dimensionality reduction methods and conventional classifiers, the hybrid classifier performs best.

Original languageEnglish
Pages (from-to)3855-3861
Number of pages7
JournalExpert Systems with Applications
Volume39
Issue number3
DOIs
Publication statusPublished - 2012 Feb 15

Fingerprint

Discriminant analysis
Support vector machines
Classifiers
Pattern recognition
Geometry
Testing

All Science Journal Classification (ASJC) codes

  • Engineering(all)
  • Computer Science Applications
  • Artificial Intelligence

Cite this

@article{364312327c9e418d930e7aef1f2c87d9,
title = "Kernel local Fisher discriminant analysis based manifold-regularized SVM model for financial distress predictions",
abstract = "Support vector machines (SVM) have demonstrated excellent performance in numerous areas of pattern recognitions. However, traditional SVM does not make efficient use of both labeled training data and unlabeled testing data. Moreover, high dimensional and nonlinear distributed data generally degrade the performance of a classifier due to the curse of dimensionality in financial distress (or bankruptcy) predictions. To address these problems, this study proposes a novel hybrid classifier which integrates Kernel local Fisher discriminant analysis (KLFDA) with a manifold-regularized SVM (MR-SVM). KLFDA is employed to find an optimal projection which maximizes the margin between data points from different classes at each local area of data manifold, while MR-SVM data-dependently warps the structure of feature space to reflect the underlying geometry of the data manifold. Compared with other dimensionality reduction methods and conventional classifiers, the hybrid classifier performs best.",
author = "Shian-Chang Huang and Yu-Cheng Tang and Lee, {Chih Wei} and Chang, {Ming Jen}",
year = "2012",
month = "2",
day = "15",
doi = "10.1016/j.eswa.2011.09.095",
language = "English",
volume = "39",
pages = "3855--3861",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier Limited",
number = "3",

}

Kernel local Fisher discriminant analysis based manifold-regularized SVM model for financial distress predictions. / Huang, Shian-Chang; Tang, Yu-Cheng; Lee, Chih Wei; Chang, Ming Jen.

In: Expert Systems with Applications, Vol. 39, No. 3, 15.02.2012, p. 3855-3861.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Kernel local Fisher discriminant analysis based manifold-regularized SVM model for financial distress predictions

AU - Huang, Shian-Chang

AU - Tang, Yu-Cheng

AU - Lee, Chih Wei

AU - Chang, Ming Jen

PY - 2012/2/15

Y1 - 2012/2/15

N2 - Support vector machines (SVM) have demonstrated excellent performance in numerous areas of pattern recognitions. However, traditional SVM does not make efficient use of both labeled training data and unlabeled testing data. Moreover, high dimensional and nonlinear distributed data generally degrade the performance of a classifier due to the curse of dimensionality in financial distress (or bankruptcy) predictions. To address these problems, this study proposes a novel hybrid classifier which integrates Kernel local Fisher discriminant analysis (KLFDA) with a manifold-regularized SVM (MR-SVM). KLFDA is employed to find an optimal projection which maximizes the margin between data points from different classes at each local area of data manifold, while MR-SVM data-dependently warps the structure of feature space to reflect the underlying geometry of the data manifold. Compared with other dimensionality reduction methods and conventional classifiers, the hybrid classifier performs best.

AB - Support vector machines (SVM) have demonstrated excellent performance in numerous areas of pattern recognitions. However, traditional SVM does not make efficient use of both labeled training data and unlabeled testing data. Moreover, high dimensional and nonlinear distributed data generally degrade the performance of a classifier due to the curse of dimensionality in financial distress (or bankruptcy) predictions. To address these problems, this study proposes a novel hybrid classifier which integrates Kernel local Fisher discriminant analysis (KLFDA) with a manifold-regularized SVM (MR-SVM). KLFDA is employed to find an optimal projection which maximizes the margin between data points from different classes at each local area of data manifold, while MR-SVM data-dependently warps the structure of feature space to reflect the underlying geometry of the data manifold. Compared with other dimensionality reduction methods and conventional classifiers, the hybrid classifier performs best.

UR - http://www.scopus.com/inward/record.url?scp=80255122809&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80255122809&partnerID=8YFLogxK

U2 - 10.1016/j.eswa.2011.09.095

DO - 10.1016/j.eswa.2011.09.095

M3 - Article

AN - SCOPUS:80255122809

VL - 39

SP - 3855

EP - 3861

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

IS - 3

ER -