Jove
Visualize
Contact Us

Related Experiment Videos

Test your surrogate data before you test for nonlinearity.

D Kugiumtzis1

  • 1Max-Planck-Institute for Physics of Complex Systems, Nöthnitzer Strasse 38, 01187 Dresden, Germany. dimitris@mpipks-dresden.mpg.de

Physical Review. E, Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics
|April 24, 2002
PubMed
Summary

Investigating nonlinear dynamics? The amplitude-adjusted Fourier transform surrogates (AAFT) algorithm has critical flaws. Iterated AAFT and nonlinear statistics offer more reliable methods for testing linear stochastic processes.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Investigation of the correlation of successive earthquakes preceding main shocks in the Greek territory.

Journal of applied statistics·2022
Same author

Direct-coupling information measure from nonuniform embedding.

Physical review. E, Statistical, nonlinear, and soft matter physics·2013
Same author

Species mobility induces synchronization in chaotic population dynamics.

Physical review. E, Statistical, nonlinear, and soft matter physics·2011
Same author

Reducing the bias of causality measures.

Physical review. E, Statistical, nonlinear, and soft matter physics·2011
Same author

Local prediction of turning points of oscillating time series.

Physical review. E, Statistical, nonlinear, and soft matter physics·2008
Same author

Statistical analysis of the extreme values of stress time series from the Portevin-Le Châtelier effect.

Physical review. E, Statistical, nonlinear, and soft matter physics·2004
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Area of Science:

  • Nonlinear dynamics
  • Time series analysis
  • Statistical hypothesis testing

Background:

  • Testing for nonlinearity in stochastic processes is crucial in many scientific fields.
  • The amplitude-adjusted Fourier transform surrogates (AAFT) algorithm is a common method for generating surrogate data to test the null hypothesis of linearity.
  • Previous studies have relied on AAFT, but its consistency has not been fully evaluated.

Purpose of the Study:

  • To investigate the consistency of surrogate data generation schemes in representing the null hypothesis of linear stochastic processes.
  • To identify and address limitations of the amplitude-adjusted Fourier transform surrogates (AAFT) algorithm.
  • To propose improved methods for nonlinearity detection in real-world data.

Main Methods:

Related Experiment Videos

  • Comparison of the standard amplitude-adjusted Fourier transform surrogates (AAFT) algorithm with the iterated AAFT method.
  • Analysis of the consistency of surrogate data generation in representing the null hypothesis.
  • Introduction and application of linear and nonlinear polynomials as discriminating statistics to detect inaccuracies.

    • Main Results:

    • The standard amplitude-adjusted Fourier transform surrogates (AAFT) algorithm exhibits significant inaccuracies, particularly in preserving linear correlations.
    • The iterated AAFT algorithm demonstrates greater consistency in representing the null hypothesis compared to the standard AAFT.
    • Inferences of nonlinearity based on the standard AAFT algorithm may be premature due to its inherent flaws.

    Conclusions:

    • The amplitude-adjusted Fourier transform surrogates (AAFT) algorithm's inaccuracies necessitate re-evaluation of previous nonlinearity findings.
    • The iterated AAFT algorithm offers a more reliable approach for generating surrogate data.
    • Employing both linear and nonlinear polynomials as discriminating statistics is recommended for robust nonlinearity detection, cautioning against the sole reliance on the AAFT algorithm.