Modelling financial time series with switching state space models

Mehdi Azzouzi, Ian T. Nabney

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The deficiencies of stationary models applied to financial time series are well documented. A special form of non-stationarity, where the underlying generator switches between (approximately) stationary regimes, seems particularly appropriate for financial markets. We use a dynamic switching (modelled by a hidden Markov model) combined with a linear dynamical system in a hybrid switching state space model (SSSM) and discuss the practical details of training such models with a variational EM algorithm due to [Ghahramani and Hilton,1998]. The performance of the SSSM is evaluated on several financial data sets and it is shown to improve on a number of existing benchmark methods.
Original languageEnglish
Title of host publicationProceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering
Place of PublicationPort Jefferson, NY
PublisherIEEE
Pages240-249
Number of pages10
ISBN (Print)0780356632
DOIs
Publication statusPublished - 1999
EventIEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering -
Duration: 1 Jan 19991 Jan 1999

Conference

ConferenceIEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering
Period1/01/991/01/99

Fingerprint

Time series
Hidden Markov models
Dynamical systems
Switches

Bibliographical note

©1999 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Keywords

  • deficiencies
  • stationary models
  • financial time series
  • non-stationarity
  • stationary regimes
  • generator switches
  • financial markets
  • hidden Markov model
  • financial data sets
  • benchmark methods

Cite this

Azzouzi, M., & Nabney, I. T. (1999). Modelling financial time series with switching state space models. In Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering (pp. 240-249). Port Jefferson, NY: IEEE. https://doi.org/10.1109/CIFER.1999.771123
Azzouzi, Mehdi ; Nabney, Ian T. / Modelling financial time series with switching state space models. Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering. Port Jefferson, NY : IEEE, 1999. pp. 240-249
@inbook{7d0880d946fb4b14a602ec818f0d976b,
title = "Modelling financial time series with switching state space models",
abstract = "The deficiencies of stationary models applied to financial time series are well documented. A special form of non-stationarity, where the underlying generator switches between (approximately) stationary regimes, seems particularly appropriate for financial markets. We use a dynamic switching (modelled by a hidden Markov model) combined with a linear dynamical system in a hybrid switching state space model (SSSM) and discuss the practical details of training such models with a variational EM algorithm due to [Ghahramani and Hilton,1998]. The performance of the SSSM is evaluated on several financial data sets and it is shown to improve on a number of existing benchmark methods.",
keywords = "deficiencies, stationary models, financial time series, non-stationarity, stationary regimes, generator switches, financial markets, hidden Markov model, financial data sets, benchmark methods",
author = "Mehdi Azzouzi and Nabney, {Ian T.}",
note = "{\circledC}1999 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.",
year = "1999",
doi = "10.1109/CIFER.1999.771123",
language = "English",
isbn = "0780356632",
pages = "240--249",
booktitle = "Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering",
publisher = "IEEE",
address = "United States",

}

Azzouzi, M & Nabney, IT 1999, Modelling financial time series with switching state space models. in Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering. IEEE, Port Jefferson, NY, pp. 240-249, IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering, 1/01/99. https://doi.org/10.1109/CIFER.1999.771123

Modelling financial time series with switching state space models. / Azzouzi, Mehdi; Nabney, Ian T.

Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering. Port Jefferson, NY : IEEE, 1999. p. 240-249.

Research output: Chapter in Book/Report/Conference proceedingChapter

TY - CHAP

T1 - Modelling financial time series with switching state space models

AU - Azzouzi, Mehdi

AU - Nabney, Ian T.

N1 - ©1999 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PY - 1999

Y1 - 1999

N2 - The deficiencies of stationary models applied to financial time series are well documented. A special form of non-stationarity, where the underlying generator switches between (approximately) stationary regimes, seems particularly appropriate for financial markets. We use a dynamic switching (modelled by a hidden Markov model) combined with a linear dynamical system in a hybrid switching state space model (SSSM) and discuss the practical details of training such models with a variational EM algorithm due to [Ghahramani and Hilton,1998]. The performance of the SSSM is evaluated on several financial data sets and it is shown to improve on a number of existing benchmark methods.

AB - The deficiencies of stationary models applied to financial time series are well documented. A special form of non-stationarity, where the underlying generator switches between (approximately) stationary regimes, seems particularly appropriate for financial markets. We use a dynamic switching (modelled by a hidden Markov model) combined with a linear dynamical system in a hybrid switching state space model (SSSM) and discuss the practical details of training such models with a variational EM algorithm due to [Ghahramani and Hilton,1998]. The performance of the SSSM is evaluated on several financial data sets and it is shown to improve on a number of existing benchmark methods.

KW - deficiencies

KW - stationary models

KW - financial time series

KW - non-stationarity

KW - stationary regimes

KW - generator switches

KW - financial markets

KW - hidden Markov model

KW - financial data sets

KW - benchmark methods

UR - http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=771123

UR - http://www.scopus.com/inward/record.url?scp=0032655279&partnerID=8YFLogxK

U2 - 10.1109/CIFER.1999.771123

DO - 10.1109/CIFER.1999.771123

M3 - Chapter

SN - 0780356632

SP - 240

EP - 249

BT - Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering

PB - IEEE

CY - Port Jefferson, NY

ER -

Azzouzi M, Nabney IT. Modelling financial time series with switching state space models. In Proceedings of the IEEE/IAFE 1999 Conference on Computational Intelligence for Financial Engineering. Port Jefferson, NY: IEEE. 1999. p. 240-249 https://doi.org/10.1109/CIFER.1999.771123