Parameter estimation of fractional Brownian motion
Hurst Parameter Estimation
This example shows how to estimate the Hurst index of a fractional Brownian motion. The example simulates 1,000 realizations of fractional Brownian motion with H=0.6. Each realization consists of 10,000 samples. At the end of the simulation, the three estimates of the Hurst index are compared.
Initialize the random number generator for repeatable results. Set the Hurst index equal to 0.6 and the length of the realizations to be 10,000.
rng default H = 0.6; len = 10000;
Generate 1,000 realizations of fractional Brownian motion and compute the estimates of the Hurst parameter.
n = 1000; Hest = zeros(n,3); for ii = 1:n fBm06 = wfbm(H,len); Hest(ii,:) = wfbmesti(fBm06); end
Compare the estimates.
subplot(3,1,1) histogram(Hest(:,1)) title("Discrete Second Derivative Estimator (DSOD)") subplot(3,1,2) histogram(Hest(:,2)) title("Wavelet Version of DSOD") subplot(3,1,3) histogram(Hest(:,3)) title("Wavelet Details Regression Estimator") xlabel("True value of the parameter H = 0.6")
X — Input signal
Input signal, specified as a vector. The signal
assumed to be a realization of fractional Brownian motion with Hurst index
hest — Fractal index estimates
Fractal index estimates, returned as a 1-by-3 vector.
hest contains three estimates of the fractal index
The first two elements of
hestare estimates based on the second derivative with the second computed in the wavelet domain.
The third estimate is based on the linear regression in loglog plot, of the variance of detail versus level.
Fractional Brownian Motion
A fractional Brownian motion (
fBm) is a
continuous-time Gaussian process depending on the Hurst parameter
0 < H
< 1. It generalizes the ordinary Brownian motion corresponding to
H = 0.5 and whose derivative is the white noise. The
fBm is self-similar in distribution and the variance of the
increments is given by
Var(fBm(t)-fBm(s)) = v |t-s|^(2H),
v is a positive constant.
This special form of the variance of the increments suggests various ways to
estimate the parameter
H. One can find in Bardet et al  a survey of
such methods. The
wfbmesti function provides three different
estimates. The first one, due to Istas and Lang , is based on
the discrete second-order derivative. The second one is a wavelet-based adaptation
and has similar properties. The third one, proposed by Flandrin , estimates
H using the slope of the loglog plot of the detail variance
versus the level. A more recent extension can be found in Abry et al .
 Abry, Patrice, Patrick Flandrin, Murad S. Taqqu, and Darryl Veitch. “Self-Similarity and Long-Range Dependence Through the Wavelet Lens.” In Theory and Applications of Long-Range Dependence, edited by P. Doukhan, G. Oppenheim, and M. Taqqu, 527–56. Birkhäuser, 2003.
 Bardet, Jean-Marc, Gabriel Lang, Georges Oppenheim, Anne Philippe, Stilian Stoev, and Murad S. Taqqu. “Semi-Parametric Estimation of the Long-Range Dependence Parameter: A Survey.” In Theory and Applications of Long-Range Dependence, edited by P. Doukhan, G. Oppenheim, and M. Taqqu, 557–77. Birkhäuser, 2003.
 Flandrin, Patrick. “Wavelet Analysis and Synthesis of Fractional Brownian Motion.” IEEE Transactions on Information Theory 38, no. 2 (March 1992): 910–17. https://doi.org/10.1109/18.119751.
 Istas, Jacques, and Gabriel Lang. “Quadratic Variations and Estimation of the Local Hölder Index of a Gaussian Process.” Annales de l’Institut Henri Poincare (B) Probability and Statistics 33, no. 4 (1997): 407–36. https://doi.org/10.1016/S0246-0203(97)80099-4.
Introduced before R2006a