Abstract
The goal of our approach is the determination of the order of the Markov process which explains the statistical dependencies of an observed time series. Our method measures the information flow of the time series indirectly via higher order cumulants considering linear and nonlinear correlations. The main point of our method, which is an extension of the method of surrogate data, is that the time series is tested against a hierarchy of nonlinear Markov processes, whose probability densities are estimated by neural networks.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
Breiman, L.: Statistics: With a View Toward Applications (Houghton Mifflin, Boston, 1973)
Deco, G., Obradovic, D.: An Information Theoretic Approach to Neural Computing(Springer, New York, 1996)
Deco, G., Schittenkopf, C., Schürmann, B.: Determining the Information Flow of Dynamical Systems from Continuous Probability Distributions. Phys. Rev. Lett. 78 (1997) 2345–2348
Eckmann, J.-P., Ruelle, D.: Ergodic Theory of Chaos and Strange Attractors. Rev. Mod. Phys. 57 (1985) 617–656
Gardiner, C.W.: Handbook of Stochastic Methods (Springer, Berlin, 1990)
Neuneier, R., Finnoff, W., Hergert, F., Ormoneit, D.: Estimation of Conditional Densities: A Comparison of Neural Network Approaches, p. 689 in Intern. Conf. on Artificial Neural Networks, ICANN (Springer, 1994)
Papoulis, A.: Probability, Random Variables, and Stochastic Processes (McGraw-Hill, Singapore, 1991)
Schittenkopf, C., Deco, G.: Testing Nonlinear Markovian Hypotheses in Dynamical Systems. Physica D 104 (1997) 61–74
Theiler, J., Eubank, S., Longtin, A., Galdrikian, B., Farmer, J.D.: Testing for Nonlinearity in Time Series: The Method of Surrogate Data. Physica D 58 (1992) 77–94
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Schittenkopf, C., Deco, G. (1997). A nonlinear markovian characterization of time series using neural networks. In: Freksa, C., Jantzen, M., Valk, R. (eds) Foundations of Computer Science. Lecture Notes in Computer Science, vol 1337. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0052117
Download citation
DOI: https://doi.org/10.1007/BFb0052117
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63746-2
Online ISBN: 978-3-540-69640-7
eBook Packages: Springer Book Archive