Are Hurst exponents estimated from short or irregular time series meaningful?

We show that several time series analysis methods that are often used for detecting self-affine fractal scaling and determining Hurst exponents in data sets may lead to spurious results when applied to short discretized data series. We show that irregularities in the series, such as jumps or spikes...

Full description

Saved in:
Bibliographic Details
Published in:Computers & geosciences 2003-11, Vol.29 (9), p.1085-1089
Main Authors: Katsev, Sergei, L’Heureux, Ivan
Format: Article
Language:English
Subjects:
Applied geophysics
Discontinuities
Earth sciences
Earth, ocean, space
Exact sciences and technology
Internal geophysics
Jumps
Power spectrum
Self-affine fractal scaling
Spikes
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We show that several time series analysis methods that are often used for detecting self-affine fractal scaling and determining Hurst exponents in data sets may lead to spurious results when applied to short discretized data series. We show that irregularities in the series, such as jumps or spikes (as are often found in geophysical data) may lead to spurious scaling and consequently to an incorrect determination of the Hurst exponent. We also illustrate the statistical error in measuring Hurst exponent in series where self-affine fractal scaling does exist. Users should be aware of these caveats when interpreting the results of short time series analysis.
ISSN:0098-3004
1873-7803
DOI:10.1016/S0098-3004(03)00105-5