Forecasting stock indices with wavelet domain kernel partial least square regressions

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

Financial time series are nonlinear and non-stationary. Most financial phenomena cannot be clearly characterized in time domain. Therefore, traditional time domain models are not very effective in financial forecasting. To address the problem, this study combines wavelet analysis with kernel partial least square (PLS) regressions for stock index forecasting. Wavelet transformation maps time domain inputs to time-frequency (or wavelet) domain, where financial characteristics can be clearly identified. Because of the high dimensionality and heavy multi-collinearity of the input data, a wavelet domain kernel PLS regressor is employed to create the most efficient subspace that maintains maximum covariance between inputs and outputs, and to perform final forecasting. Empirical results demonstrate that the proposed model outperforms traditional neural networks, support vector machines, GARCH models, and has significantly reduced the forecasting errors.

Original languageEnglish
Pages (from-to)5433-5443
Number of pages11
JournalApplied Soft Computing Journal
Volume11
Issue number8
DOIs
Publication statusPublished - 2011 Dec 1

Fingerprint

Wavelet analysis
Support vector machines
Time series
Neural networks

All Science Journal Classification (ASJC) codes

  • Software

Cite this

@article{989a277de336436d83bc212784e485ab,
title = "Forecasting stock indices with wavelet domain kernel partial least square regressions",
abstract = "Financial time series are nonlinear and non-stationary. Most financial phenomena cannot be clearly characterized in time domain. Therefore, traditional time domain models are not very effective in financial forecasting. To address the problem, this study combines wavelet analysis with kernel partial least square (PLS) regressions for stock index forecasting. Wavelet transformation maps time domain inputs to time-frequency (or wavelet) domain, where financial characteristics can be clearly identified. Because of the high dimensionality and heavy multi-collinearity of the input data, a wavelet domain kernel PLS regressor is employed to create the most efficient subspace that maintains maximum covariance between inputs and outputs, and to perform final forecasting. Empirical results demonstrate that the proposed model outperforms traditional neural networks, support vector machines, GARCH models, and has significantly reduced the forecasting errors.",
author = "Shian-Chang Huang",
year = "2011",
month = "12",
day = "1",
doi = "10.1016/j.asoc.2011.05.015",
language = "English",
volume = "11",
pages = "5433--5443",
journal = "Applied Soft Computing",
issn = "1568-4946",
publisher = "Elsevier BV",
number = "8",

}

Forecasting stock indices with wavelet domain kernel partial least square regressions. / Huang, Shian-Chang.

In: Applied Soft Computing Journal, Vol. 11, No. 8, 01.12.2011, p. 5433-5443.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Forecasting stock indices with wavelet domain kernel partial least square regressions

AU - Huang, Shian-Chang

PY - 2011/12/1

Y1 - 2011/12/1

N2 - Financial time series are nonlinear and non-stationary. Most financial phenomena cannot be clearly characterized in time domain. Therefore, traditional time domain models are not very effective in financial forecasting. To address the problem, this study combines wavelet analysis with kernel partial least square (PLS) regressions for stock index forecasting. Wavelet transformation maps time domain inputs to time-frequency (or wavelet) domain, where financial characteristics can be clearly identified. Because of the high dimensionality and heavy multi-collinearity of the input data, a wavelet domain kernel PLS regressor is employed to create the most efficient subspace that maintains maximum covariance between inputs and outputs, and to perform final forecasting. Empirical results demonstrate that the proposed model outperforms traditional neural networks, support vector machines, GARCH models, and has significantly reduced the forecasting errors.

AB - Financial time series are nonlinear and non-stationary. Most financial phenomena cannot be clearly characterized in time domain. Therefore, traditional time domain models are not very effective in financial forecasting. To address the problem, this study combines wavelet analysis with kernel partial least square (PLS) regressions for stock index forecasting. Wavelet transformation maps time domain inputs to time-frequency (or wavelet) domain, where financial characteristics can be clearly identified. Because of the high dimensionality and heavy multi-collinearity of the input data, a wavelet domain kernel PLS regressor is employed to create the most efficient subspace that maintains maximum covariance between inputs and outputs, and to perform final forecasting. Empirical results demonstrate that the proposed model outperforms traditional neural networks, support vector machines, GARCH models, and has significantly reduced the forecasting errors.

UR - http://www.scopus.com/inward/record.url?scp=80053567692&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053567692&partnerID=8YFLogxK

U2 - 10.1016/j.asoc.2011.05.015

DO - 10.1016/j.asoc.2011.05.015

M3 - Article

AN - SCOPUS:80053567692

VL - 11

SP - 5433

EP - 5443

JO - Applied Soft Computing

JF - Applied Soft Computing

SN - 1568-4946

IS - 8

ER -