Direct calculation of metric entropy from time series (Q1208857)

Some time ago Kolmogorov introduced the concept of metric entropy (which can also the thought of as a geometrical entropy) which is useful in the analysis of dynamical systems, and mainly provides a characterization of chaotic dynamics. The major problem is that it is very hard to calculate this entropy directly by using the original definition. It is the contribution of the present paper to show that this can be done with fine partitions and long sequence lengths by using partitioning and sequence compression techniques. The method is then tested for the attractors of Henon, Ikeda, Rossler, Lorentz, and Mackey-Glass, respectively.