Techniques for multifractal spectrum estimation in financial time series
Free (open access)
Volume 10 (2015), Issue 3
261 - 266
P. JIZBA & J. KORBEL
We show that a multifractal analysis offers a new and potentially promising avenue for quantifying the complexity of various time series. In particular, we compare the most common techniques used for multifractal scaling exponents estimation. This is done from both a theoretical and phenomenological point of view. In our discussion we speciﬁcally focus on methods based on estimation of Rényi entropy, which provide a powerful tool especially in the presence of heavy-tailed data. As a testbed for the applicability of above multifractal methods we use various real ﬁnancial datasets, including both daily and high-frequency data.
multifractal spectrum, Rényi entropy, time series