Detrended Fluctuation Analysis — Long-Range Correlations

Signal Parameters

Signal type: fBm Target Hurst H: 0.75 Signal length: 2048 DFA order: 1 Noise: 0.1

Estimated α (DFA):—

True H:—

Class:—

Variance:—

DFA algorithm: 1) Integrate signal, 2) divide into windows of size n, 3) detrend each window (polynomial fit), 4) compute RMS fluctuation F(n), 5) fit F(n)~nᵅ.

α=0.5: white noise. α=1: 1/f (pink). α=1.5: Brownian. α>0.5: persistent. α<0.5: anti-persistent.

Detrended Fluctuation Analysis (DFA)

Signal Parameters