Improvements of boxed correlation sum

JuliaDynamics / ChaosTools.jl

Tools for the exploration of chaos and nonlinear dynamics

https://juliadynamics.github.io/DynamicalSystemsDocs.jl/chaostools/stable/

MIT License

189 stars 36 forks source link

Improvements of boxed correlation sum #207

Closed Datseris closed 3 years ago

Datseris commented 3 years ago

@apbraun let's do all necessary improvements in this PR.

Datseris commented 3 years ago

@apbraun can you please remind me: these are all the improvements we wanted to do, right? I can merge away?

apbraun commented 3 years ago

Yes these are all improvements. We spoke about adjusting the _autoprism algorithm to the Theiler version and catch r0 < ε0 in the boxed_correlationsum, I also spotted another bug in boxed_correlationsum that I fixed, so from my part we're ready to go.

Datseris commented 3 years ago

@apbraun please give me some help here. In our paper you said that "Bueno-Orovio and Pérez-García chose P=2 for D>2" where I assume that P stands for the Prism dimension. Here we don't do that, but we use I think an older approach by Theiler. Which one should we use in the end...?