MULTIFRACTAL DETRENDED FLUCTUATION ANALYSIS AS THE ESTIMATOR OF LONG-RANGE DEPENDENCE

Detection method of long-range dependences based on multifractal analysis in time series is proposed. A short description of multifractal analysis and estimator construction (based on Multifractal Detrended Fluctuation Analysis) are given. This method gives accurate results when applied to large scale analysis of fractional Brownian motions (fBm) and describes consistently the mixture of two fBm processes. Finally, this method has been applied to series corresponding K+ ionic current through the cellular membrane. For times shorter than 1 sec, a similarity between ionic K+ current and the mixture of two antipersistent processes has been found with self-similarity parameter of this mixture less than 0.30.