Abstract
Non-stationary time series are extremely challenging to model. We propose a Bayesian mixture model framework for obtaining time varying parameters for a dynamic linear model. We discuss on-line estimation of time varying DLM parameters by means of a dynamic mixture model composed of constant parameter DLMs. For time series with low signal-to-noise ratios, we propose a novel method of constructing model priors. We calculate model likelihoods by comparing forecast distributions with observed values. We utilize computationally efficient moment matching Gaussians to approximate exact mixtures of path dependent posterior densities. The effectiveness of our approach is illustrated by extracting insightful time varying parameters for an ETF returns model in a period spanning the 2008 financial crisis; and, by demonstrating the superior performance in a statistical arbitrage application.
Access provided by Autonomous University of Puebla. Download to read the full chapter text
Chapter PDF
Similar content being viewed by others
References
Johnson, R., Wichern, D.: Applied Multivariate Statistical Analysis. Prentice Hall, Upper Saddle River (2002)
West, M., Harrison, J.: Bayesian Forecasting and Dynamic Models. Springer-Verlag New York, Inc., New York (1997)
Kalman, R., et al.: A new approach to linear filtering and prediction problems. Journal of Basic Engineering 82, 35–45 (1960)
Kalaba, R., Tesfatsion, L.: A multicriteria approach to model specification and estimation. Computational Statistics & Data Analysis 21, 193–214 (1996)
Minka, T.: From hidden Markov models to linear dynamical systems. Technical Report 531, Vision and Modeling Group of Media Lab, MIT (1999)
Bishop, C.: Pattern Recognition and Machine Learning (Information Science and Statistics). Springer Science+Business Media, LLC, New York (2006)
Hamilton, J.: Time series analysis. Princeton University Press, Princeton (1994)
Montana, G., Triantafyllopoulos, K., Tsagaris, T.: Flexible least squares for temporal data mining and statistical arbitrage. Expert Systems with Applications 36, 2819–2830 (2009)
Crassidis, J., Cheng, Y.: Generalized Multiple-Model Adaptive Estimation Using an Autocorrelation Approach. In: 2006 9th International Conference on Information Fusion, pp. 1–8. IEEE (2007)
Ghahramani, Z., Hinton, G.: Parameter estimation for linear dynamical systems. Technical Report CRG-TR-96-2, University of Toronto (1996)
Valpola, H., Harva, M., Karhunen, J.: Hierarchical models of variance sources. Signal Processing 84, 267–282 (2004)
Sarkka, S., Nummenmaa, A.: Recursive noise adaptive Kalman filtering by variational Bayesian approximations. IEEE Transactions on Automatic Control 54, 596–600 (2009)
Minka, T.P.: Bayesian inference in dynamic models: an overview (2007), http://research.microsoft.com
Chen, R., Liu, J.: Mixture Kalman filters. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 62, 493–508 (2000)
Bar-Shalom, Y., Li, X., Kirubarajan, T., Wiley, J.: Estimation with applications to tracking and navigation. John Wiley & Sons, Inc. (2001)
PDR Services LLC: Prospectus. SPDR S&P 500 ETF (2010), https://www.spdrs.com
Rydex Distributors, LLC: Prospectus. Rydex S&P Equal Weight ETF (2010), http://www.rydex-sgi.com/
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Keane, K.R., Corso, J.J. (2013). Inferring Time Varying Dynamic Linear Model Parameters from a Mixture Model. In: Latorre Carmona, P., Sánchez, J., Fred, A. (eds) Pattern Recognition - Applications and Methods. Advances in Intelligent Systems and Computing, vol 204. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-36530-0_4
Download citation
DOI: https://doi.org/10.1007/978-3-642-36530-0_4
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-36529-4
Online ISBN: 978-3-642-36530-0
eBook Packages: EngineeringEngineering (R0)