JuliaDynamics / ChaosTools.jl

Tools for the exploration of chaos and nonlinear dynamics
https://juliadynamics.github.io/DynamicalSystemsDocs.jl/chaostools/stable/
MIT License
187 stars 35 forks source link

Basins labels for supervised clustering method #265

Closed KalelR closed 2 years ago

KalelR commented 2 years ago

Currently, basins_of_attraction fails for non-number labels given in templates since basins is initialized as an Ìnt` matrix.

A proposal would be to keep basins as an Int matrix for the computations (via basins_fractions) and then replace its number values by the arbitrary non-number labels from `templates' later. Specifically, we would change this line

vec(basins) .= vec(labels)

when labels is not an Integer. This would be very similar to how we replace labels in cluster_features_distances.