Time series analysis by using permutation entropy and statistical complexity

The starting point to study many dynamical phenomena in nature is a
set of measurements of some representative variable of interest at
discrete time intervals, i.e. a \"black box\" time series. When the
underlying dynamics is simple, traditional and standard tools, like
Fourier tranform, can be used to make inferences and characterize the
behavior. However, for complex dynamics more sophisticated approaches
are necessary. In this seminar I will introduce two novel information
theory quantifiers: permutation entropy and statistical complexity.
Useful information about the system\'s dynamics can be derived from
them. Moreover, they can be easily estimated and are robust under the
presence of additive noise. Several applications, from the analysis of
numerical and experimental time series, will be shown.