MFDFA: Efficient Multifractal Detrended Fluctuation Analysis in Python

TL;DR

This paper introduces a fast, user-friendly Python library for Multifractal Detrended Fluctuation Analysis (MFDFA), enabling efficient analysis of time series variability and correlations with multi-threaded processing.

Contribution

The paper presents an efficient, easy-to-use Python implementation of MFDFA that includes common extensions and leverages multi-threading for rapid computations.

Findings

Significantly faster MFDFA computations with multi-threading.

Inclusion of common extensions like empirical mode decomposition.

Enhanced usability for analyzing time series variability.

Abstract

Multifractal detrended fluctuation analysis (MFDFA) has become a central method to characterise the variability and uncertainty in empiric time series. Extracting the fluctuations on different temporal scales allows quantifying the strength and correlations in the underlying stochastic properties, their scaling behaviour, as well as the level of fractality. Several extensions to the fundamental method have been developed over the years, vastly enhancing the applicability of MFDFA, e.g. empirical mode decomposition for the study of long-range correlations and persistence. In this article we introduce an efficient, easy-to-use python library for MFDFA, incorporating the most common extensions and harnessing the most of multi-threaded processing for very fast calculations.