Wavelet-based discrimination of isolated singularities masquerading as multifractals in detrended fluctuation analyses

Abstract

The robustness of two widespread multifractal analysis methods, one based on detrended fluctuation analysis and one on wavelet leaders, is discussed in the context of time-series containing non-uniform structures with only isolated singularities. Signals generated by simulated and experimentally-realized chaos generators, together with synthetic data addressing particular aspects, are taken into consideration. The results reveal essential limitations affecting the ability of both methods to correctly infer the non-multifractal nature of signals devoid of a cascade-like hierarchy of singularities. Namely, signals harboring only isolated singularities are found to artefactually give rise to broad multifractal spectra, resembling those expected in the presence of a well-developed underlying multifractal structure. Hence, there is a real risk of incorrectly inferring multifractality due to isolated singularities. The careful consideration of local scaling properties and the distribution of Hölder exponent obtained, for example, through wavelet analysis, is indispensable for rigorously assessing the presence or absence of multifractality.

To appear in Nonlinear Dynamics

References in corpus (7)

• Detrending moving average algorithm for multifractals
• Detecting and interpreting distortions in hierarchical organization of complex time series
• Generalized Hurst exponent and multifractal function of original and translated texts mapped into frequency and length time series
• Long-range correlation and multifractality in Bach's Inventions pitches
• Long-term correlations and multifractal nature in the intertrade durations of a liquid Chinese stock and its warrant
• Boundary criticality and multifractality at the 2D spin quantum Hall transition
• Uncovering latent singularities from multifractal scaling laws in mixed asymptotic regime. Application to turbulence