Comparing permutation entropy functions to detect structural changes in time series

Abstract

Entropy can be taken as a measure of the complex dynamical behavior. In this paper, we consider different entropy functions and the permutation symbolic dynamics and we apply them to find structural changes in time series. We analyze what entropy functions are more suitable to show changes in simulated time series where the structural changes are know. Applications to seismic real data and economic data series are shown to illustrate how this type of tools can be used.