DYNAMICALLY MIXING DYNAMIC LINEAR MODELS WITH APPLICATIONS IN FINANCE

Kevin R. Keane, Jason J. Corso

2012

Abstract

Time varying model parameters offer tremendous flexibility while requiring more sophisticated learning methods. 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. We conclude by demonstrating the superior performance of time varying mixture models against constant parameter DLMs in a statistical arbitrage application.

References

  1. Bar-Shalom, Y., Li, X., Kirubarajan, T., and Wiley, J. (2001). Estimation with applications to tracking and navigation. John Wiley & Sons, Inc.
  2. Bishop, C. (2006). Pattern Recognition and Machine Learning (Information Science and Statistics). Springer Science+Business Media, LLC. New York, NY, USA.
  3. Chen, R. and Liu, J. (2000). Mixture Kalman filters. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 62(3):493-508.
  4. Crassidis, J. and Cheng, Y. (2007). Generalized MultipleModel Adaptive Estimation Using an Autocorrelation Approach. In Information Fusion, 2006 9th International Conference on, pages 1-8. IEEE.
  5. Ghahramani, Z. and Hinton, G. (1996). Parameter estimation for linear dynamical systems. Technical Report CRG-TR-96-2, University of Toronto.
  6. Hamilton, J. (1994). Time series analysis. Princeton University Press: Princeton, NJ, USA.
  7. Johnson, R. and Wichern, D. (2002). Applied Multivariate Statistical Analysis. Prentice Hall: Upper Saddle River, NJ, USA.
  8. Kalaba, R. and Tesfatsion, L. (1996). A multicriteria approach to model specification and estimation. Computational Statistics & Data Analysis, 21(2):193-214.
  9. Kalman, R. et al. (1960). A new approach to linear filtering and prediction problems. Journal of Basic Engineering, 82(1):35-45.
  10. Minka, T. (1999). From hidden Markov models to linear dynamical systems. Technical Report 531, Vision and Modeling Group of Media Lab, MIT.
  11. Minka, T.P. (2007). Bayesian inference in dynamic models: an overview. http://research.microsoft.com.
  12. Montana, G., Triantafyllopoulos, K., and Tsagaris, T. (2009). Flexible least squares for temporal data mining and statistical arbitrage. Expert Systems with Applications, 36(2):2819-2830.
  13. PDR Services LLC (2010). Prospectus. SPDR S&P 500 ETF. https://www.spdrs.com.
  14. Rydex Distributors, LLC (2010). Prospectus. Rydex S&P Equal Weight ETF. http://www.rydex-sgi.com/.
  15. Sarkka, S. and Nummenmaa, A. (2009). Recursive noise adaptive Kalman filtering by variational Bayesian approximations. Automatic Control, IEEE Transactions on, 54(3):596-600.
  16. Valpola, H., Harva, M., and Karhunen, J. (2004). Hierarchical models of variance sources. Signal Processing, 84(2):267-282.
  17. West, M. and Harrison, J. (1997). Bayesian Forecasting and Dynamic Models. Springer-Verlag New York, Inc. New York, NY, USA.
  18. Yelland, P. and Lee, E. (2003). Forecasting product sales with dynamic linear mixture models. Technical Report SMLI TR-2003-122, Sun Microsystems, Inc.
Download


Paper Citation


in Harvard Style

R. Keane K. and J. Corso J. (2012). DYNAMICALLY MIXING DYNAMIC LINEAR MODELS WITH APPLICATIONS IN FINANCE . In Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods - Volume 2: ICPRAM, ISBN 978-989-8425-99-7, pages 295-302. DOI: 10.5220/0003712602950302


in Bibtex Style

@conference{icpram12,
author={Kevin R. Keane and Jason J. Corso},
title={DYNAMICALLY MIXING DYNAMIC LINEAR MODELS WITH APPLICATIONS IN FINANCE},
booktitle={Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods - Volume 2: ICPRAM,},
year={2012},
pages={295-302},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0003712602950302},
isbn={978-989-8425-99-7},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 1st International Conference on Pattern Recognition Applications and Methods - Volume 2: ICPRAM,
TI - DYNAMICALLY MIXING DYNAMIC LINEAR MODELS WITH APPLICATIONS IN FINANCE
SN - 978-989-8425-99-7
AU - R. Keane K.
AU - J. Corso J.
PY - 2012
SP - 295
EP - 302
DO - 10.5220/0003712602950302