Fast Ordinal Pattern and Permutation Entropy computation

Fast methods for computing Shannon Entropies regarding ordinal patterns and words

This package provides a collection of methods for computing entropies from symbolical dynamics.
There are methods to compute all ordinal patterns of a given Data Vector in different representations, namely the permutation and inversion vector representation as well as our new approach, the binary vector representation.
For each representation, we give two algorithms, one that simply iterates through the Data Vector and computes the patterns sequencially and one that takes benefit of overlapping ordinal information of the preceeding patterns and computes them successively.
The latter approach gives a substancial speed improvement.

We use the newly introduced binary vector representation to compute a whole set of Entropies: Let D<=52 be given, then we compute all empirical Shannon entropies H(d,m) for ordinal words of order d and length m where d+m-1<=D.
H(d,1)/d gives the well known empirical permutation entropy.

We offer one entropy computing function that computes the ordinal patterns itself as well as a function that takes precomputed patterns as an input.

Our methods allow higher values for the pattern order than many other programms.