Divergence Measure Between Chaotic Attractors

Abstract

We propose a measure of divergence of probability distributions for quantifying the dissimilarity of two chaotic attractors. This measure is defined in terms of a generalized entropy. We illustrate our procedure by considering the effect of additive noise in the well known Hénon attractor. Comparison of two Hénon attractors for slighly different parameter values, has shown that the divergence has complex scaling structure. Finally, we show how our approach allows to detect non-stationary events in a time series.

9 pages, 6 figures

References in corpus (1)

• Detecting and analysing nonstationarity in a time series with nonlinear cross-predictions