Abstract
Most real world time series have transient behaviours and are non-stationary. They exhibit different type of non-stationarities, such as trends, cycles, random-walking, and generally exhibit strong intermittency. Therefore local stochastic characteristics of time series, such as the drift and diffusion coefficients, as well as the jump rate and jump amplitude, will provide very important information for understanding and quantifying “real time” variability of time series. For diffusive processes the systems have a longer memory and a higher correlation time scale and, therefore, one expects the stochastic features of dynamics to change slowly. In contrast, a rapid change of dynamics with jumps will cause strong ramp events (abrupt changes) in small time scales. Beside nonstationarity, in real world data, such as stock market index, pattern of cosmic background radiation, genomic data, etc., there is unique trajectory. Therefore, we require analysing techniques to estimate local stochastic behaviour of time series, which will be applicable to stationary and nonstationary time series and those with unique trajectory. This chapter contains technical aspects of the approach for real-time estimating of the Kramers–Moyal (KM) coefficients, and the drift and diffusion coefficients, as well as jump contributions for time series. We present the Kernel method (Nadaraya-Watson estimator) to estimate the time-dependent KM coefficients, which can be used in analysing stationary and nonstationary time series. We will also provide the details of estimating the KM coefficients in the presence of microstructure (measurement) noise, and show how the statistical properties of the noise can be determined from vanishing \(\tau \) limit behaviour of the KM conditional moments.
[Type][CrossLinking]The original version of this chapter was revised: Belated correction has been incorporated. The correction to this chapter is available at https://doi.org/10.1007/978-3-030-18472-8_24
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
To have an accurate estimation for marginal probability density, if one needs n data points with bin size \(\delta _1\), for two point joint or conditional probability density will need approximately order 10n data points. This increases the bin size to \(\delta _2=2.15 \delta _1\). The bin size for marginal density will be \( \delta _1 \le 2.6 (IQ) n^{-1/3}\) (found by Terrell), where (IQ) is the inter-quartile range of data set. The inter-quartile range is given by \((IQ) = Q_3 - Q_1\). To determine the values of \(Q_1\) and \(Q_3\), find the median of whole data set, separate data to “above” and “below” the median, find median for each new data set, name the median of “above” and “below” sets as \(Q_3\) and \(Q_1\), respectively. Typical bin numbers for normalised time series with \(10^6\) data points, will be \(M_1\simeq 121\) and \(M_2\simeq 55\) bins for estimation of marginal and conditional PDFs. In practice we use an odd number of bins. In Kramers–Moyal coefficients we need to estimate conditional PDF \(p(x',t+\tau |x,t)\) therefore can use an approximate bin size \(\delta _2\) in the analysis. For more detailed analysis use the method proposed by Knuth to obtain optimal bin numbers for joint probability distribution.
For stationary time series with n data points, (1) estimate the Kramers–Moyal coefficients via binning of state variable and using the Nadaraya-Watson estimator with different bandwidth h. (2) Select bandwidth \(h^*\) in Nadaraya-Watson estimator, in such a way that the estimated KM coefficients from two approaches converge. This gives bandwidth \(h^*\), which is cross-validated with binning method. (3) Now apply kernel estimator with obtained bandwidth to original data and compute time dependent KM coefficients.
References
E.A. Nadaraya, On estimating regression, theory of probability and its applications. Theory Probab. Appl. 9, 141 (1964)
G.S. Watson, Smooth regression analysis. Sankhya: Indian J. Stat., Ser. A 26, 359 (1964)
H.J. Bierens, Kernel estimators of regression functions, in Advances in Econometrics, Fifth World Congress of the Econometric Society, ed. by T.F. Bewley, vol. I of Econometric Society Monographs, chap. 3, pp. 99-144. (Cambridge University Press, Cambridge, 1987)
B.W. Silverman, Density Estimation (Chapman and Hall, London, 1986)
C. Honisch, Analysis of Complex Systems: From Stochastic Time Series to Pattern Formation in Microscopic Fluidic Films (Dissertation, University of Münster) (Westfalen, 2014)
M. Köhler, A. Schindler, S. Sperlich, A review and comparison of bandwidth selection methods for kernel regression. Int. Stat. Rev. 82, 243 (2014)
J. Gao, I. Gijbels, Bandwidth selection in nonparametric kernel testing. J. Amer. Statist. Assoc. 103, 1584–1594 (2008)
G.R. Terrell, The maximal smoothing principle in density estimation. J. Am. Stat. Assoc. 85, 470 (1990)
S. Ghosh, Kernel Smoothing: Principles, Methods And Applications (Wiley, New York, 2017)
K.H. Knuth, Optimal data-based binning for histograms, ArXiv Physics e-prints (May, 2006) [physics/0605197]
H. Kantz, T. Schreiber, Nonlinear Time Series Analysis (Cambridge University Press, Cambridge, 2003)
E.J. Kostelich, T. Schreiber, Phys. Rev. E 48, 1752 (1993)
J.P.M. Heald, J. Stark, Phys. Rev. Lett. 84, 2366 (2000)
F. Böttcher, J. Peinke, D. Kleinhans, R. Friedrich, P.G. Lind, M. Haase, Phys. Rev. Lett. 97, 090603 (2006)
M. Siefert, A. Kittel, R. Friedrich, J. Peinke, Europhys. Lett. 61, 466 (2003)
K. Lehnertz, L. Zabawa, M. Reza Rahimi Tabar, New J. Phys. 20, 113043 (2018)
P.G. Lind, A. Mora, J.A.C. Gallas, M. Haase, Int. J. Bifurc. Chaos 17, 3461 (2007)
C. Renner, J. Peinke, R. Friedrich, Phys. A 298, 499 (2001)
M. Anvari, K. Lehnertz, M.R. Rahimi Tabar, J. Peinke, Sci. Rep. 6, 35435 (2016)
Author information
Authors and Affiliations
Corresponding author
Problems
Problems
17.1
Nadaraya-Watson Estimator, Ornstein–Uhlenbeck Process with Jumps
Use Nadaraya-Watson estimator (use a Gaussian kernel with bandwidth \(h\simeq 0.1\)) to estimate time dependent Kramers–Moyal coefficients and reproduce the results presented in Fig. 17.1. The drift and diffusion coefficients as well as jump amplitude and jump rate are given in the Sects. 17.1.1 and 21.2.1. Use the Euler-Maruyama scheme to integrate related jump-diffusion dynamical equation.
17.2
Local Drift and Diffusion of the Minimal Market Model
The minimal market model is given by a (Itô) Langevin dynamics with time dependent drift a(x, t) and multiplicative b(x, t) as,
with scaling parameter \(\alpha > 0\) and growth rate \(\eta >0\).
(a) Use the Euler-Maruyama scheme to approximate this Langevin equation and generate \(10^5\) data points of \(x(t_n)\) with \(x(t_0)= x_0=1\), \(dt=0.001\), \(\alpha =0.2\) and \(\eta =0.01\).
(b) Use the Nadaraya-Watson estimator with Gaussian kernel (with different bandwidth \(h \in \{0.01,0.1,0.6\}\)) to estimate time dependence of the functions a and b. The first and second order Kramers–Moyal coefficients are given by \(D^{(1)} (x,t) = a(x,t)\) and \(b(x,t) = \sqrt{2D^{(2)} (x,t)}\).
17.3
Microstructure Noise
Derive the third- and fourth-order (\(j=3,4\)) Kramers–Moyal conditional moments \(K^{(j)}(y, \tau )\) in terms of \(K^{(j)}(x, \tau )\) and statistical moments of noise \(\eta (t)\) in the weak noise limit. Assume that the noise to be uncorrelated with x(t), that it has finite statistical moments \(\langle \eta ^{n} \rangle \) for \(n=2,\cdots \).
17.4
Microstructure Noise
Consider a Ornstein–Uhlenbeck process x(t) with drift and diffusion coefficients as \(D^{(1)} (x,t) = -10x\) and \(D^{(2)} (x,t) = 1\), respectively, and synthesis a time series with \(10^6\) data points (with \(x(t=0)=0\) and \(dt=0.01\)).
Now generate a new time series with microstructure noise as,
where \(\sigma =0.01\) and \(\zeta (t)\) is uncorrelated zero mean, unit variance Gaussian noise \(\zeta \sim \mathcal{N} (0,\sigma _{\zeta } ^2=1)\).
(a) Estimate second order Kramers–Moyal conditional moment for process y(t), and using the method presented in Sect. 17.2, determine \(\sigma \).
(b) Estimate drift and diffusion coefficients of original process x(t).
(c) Find the explicit expressions for \(\gamma _{1,2}(y)\) in terms of y, \(\sigma \) and variance of x [14].
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Tabar, M.R.R. (2019). The Kramers–Moyal Coefficients of Non-stationary Time Series and in the Presence of Microstructure (Measurement) Noise. In: Analysis and Data-Based Reconstruction of Complex Nonlinear Dynamical Systems. Understanding Complex Systems. Springer, Cham. https://doi.org/10.1007/978-3-030-18472-8_17
Download citation
DOI: https://doi.org/10.1007/978-3-030-18472-8_17
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-18471-1
Online ISBN: 978-3-030-18472-8
eBook Packages: Physics and AstronomyPhysics and Astronomy (R0)