Modelling financial time series with switching state space models

Author

Azzouzi, Mehdi ; Nabney, Ian T.

Author_Institution

Neural Comput. Res. Group, Aston Univ., Birmingham, UK

fYear

1999

fDate

1999

Firstpage

240

Lastpage

249

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 Z. Ghahramani and G.E. Hinton (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

financial data processing; hidden Markov models; state-space methods; time series; variational techniques; SSSM; benchmark methods; dynamic switching; financial data sets; financial markets; financial time series modelling; hidden Markov model; hybrid switching state space model; linear dynamical system; non-stationarity; stationary models; stationary regimes; switching state space models; underlying generator; variational EM algorithm; Control engineering; Data engineering; Econometrics; Economic forecasting; Exchange rates; Hidden Markov models; Jacobian matrices; Predictive models; State-space methods; Switches;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence for Financial Engineering, 1999. (CIFEr) Proceedings of the IEEE/IAFE 1999 Conference on

Conference_Location

New York, NY

Print_ISBN

0-7803-5663-2

Type

conf

DOI

10.1109/CIFER.1999.771123

Filename

771123