Sparse notes:
-
expand MultiDataset to such that it has named modalities, instead of Integer names.
-
There's a warning in MultiDataset->DataFrame conversion?
-
TODO: hide dimensionality of the instances:
const AbstractDimensionalDataset{T<:Number,D} = AbstractVector{<:AbstractArray{T,D}} -
enforce class (and regressor) variables not be part any modality
-
Integration with ScientificTypes.jl
-
Export as .h5mu, .h5ad
-
Add parameter skipextremes etc.
-
Fix allfeatvalues
-
Uniform discretaization interface
Add parameters to listrules:
- max_natoms
- max_consequent_natoms
- max_antecedent_natoms
- max_nconjuncts
- max_consequent_nconjuncts
- max_antecedent_nconjuncts