|
Activity Number:
|
566
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #304396 |
|
Title:
|
Assessing Time-Changing Hurst Exponent and Variance in Multifractional Brownian Motion
|
|
Author(s):
|
Kichun S. Lee*+ and Brani Vidakovic
|
|
Companies:
|
Georgia Institute of Technology and Georgia Tech/Emory University
|
|
Address:
|
313 Ferst Dr. , Atlanta, GA, 30332,
|
|
Keywords:
|
multifractional Brownian motion ; Hurst exponent ; EEG signals
|
|
Abstract:
|
Hurst exponent and variance are two quantities that often characterize real-life high-frequency observations. We develop and test the method for joint estimation of a time changing Hurst exponent $H(t)$ and constant scale parameter $C$ in multifractional Brownian motion based on the asymptotic behavior of the local k-th variation of its sampled paths. This work builds on research of Coeurjolly and Istas and Lang and provides a stable, simultaneous estimator of both parameters. We also discuss the asymptotic properties of joint estimators and stability of computations that use adapted wavelet filters. The procedure is evaluated on simulated data and applied on detection of vasospasm in multichannel EEG data.
|
