Comparison study of DFA and DMA methods in analysis of autocorrelations in time series

Original · EN

Statistics of the Hurst scaling exponents calculated with the use of two methods: recently introduced Detrended Moving Average Analysis(DMA) and Detrended Fluctuation Analysis (DFA)are compared. Analysis is done for artificial stochastic Brownian time series of various length and reveals interesting statistical relationships between two methods. Good agreement between DFA and DMA techniques is found for long time series L 10⁵, however for shorter series we observe that two methods give different results with no systematic relation between them. It is shown that, on the average, DMA method overestimates the Hurst exponent comparing it with DFA technique.