Related Experiment Videos
Effect of nonlinear filters on detrended fluctuation analysis
Zhi Chen1, Kun Hu, Pedro Carpena
1Center for Polymer Studies and Department of Physics, Boston University, Boston, Massachusetts 02215, USA.
Physical Review. E, Statistical, Nonlinear, and Soft Matter Physics
|February 9, 2005
Summary
Detrended fluctuation analysis (DFA) reveals how linear and nonlinear transformations impact signal correlations. Nonlinear filters alter scaling properties, unlike linear filters, depending on signal characteristics and filter parameters.
Area of Science:
- Complex Systems Analysis
- Signal Processing
- Statistical Physics
Background:
- Investigating complex systems requires understanding how measurable outputs relate to underlying quantities.
- Signal transformations, both linear and nonlinear, can obscure or alter these relationships.
- Detrended fluctuation analysis (DFA) is a key method for quantifying power-law correlations in nonstationary signals.
Purpose of the Study:
- To investigate the effects of linear and nonlinear transformations on signal correlation and scaling properties.
- To analyze the performance of DFA when applied to transformed signals and common analytic functions.
Main Methods:
- Applied linear (y=ax+b), nonlinear polynomial (y=ax^k), and nonlinear logarithmic (y=log(x+Δ)) filters to signals.
- Compared correlation and scaling properties before and after signal transformation using DFA.
- Examined DFA's application to exponential, logarithmic, and power-law functions to assess apparent scaling.
Main Results:
- Linear filters did not alter correlation properties.
- Nonlinear filters' effects depended on original signal correlations, polynomial power (k), and logarithmic offset (Δ).
- DFA curves showed identical slopes for exponential, logarithmic, and power-law functions over a range of parameter 'a'.
Conclusions:
- Nonlinear transformations significantly impact signal correlation and scaling properties, requiring careful consideration in complex system analysis.
- DFA performance is robust across different analytic functions under specific parameter conditions.
- Understanding transformation effects is crucial for accurate interpretation of dynamical properties in physical and physiological systems.