Art
@nalimilan ready for another review round.
`transform` & `reconstruct` are internal function of StatsBase. They are not exported so do not try to override them. `MultivariateStats` reexports `fit` & `predict` from the StatsAPI. `predict` call does...
Yes, we should definitely put `reconstruct` into StatsAPI.
@tpdsantos Thanks for the effort. However, can we make these changes gradually. I will not be able to review PR in this form. Some of the things that I noticed:...
> I wanted to give the user the ability to easily change the mutation functions BEFORE entering the `ga` function. My main goal was to compile all that I could...
> What do you think the branch needs to be accepted? - Start from introducing smaller PR that prepare layout for large changes. - Make unrelated changes into a single...
> If an individual is, for instance, composed by an IntegerGene and two FloatGenes, you would need to pass two types of mutation functions. Exactly, none of the package mutation...
I'm rewriting the package interface to provide more modular solution for existing and new implementations, see #49. It's very much follows the [Optim.jl](https://github.com/JuliaNLSolvers/Optim.jl) package design. The new API will provide...
This package performs an objective function minimization by maximizing its inverse. Shape your problem to fit the package behavior.
> Maybe maxdepth is no longer respected? Indeed. I wrote very basic tree mutation & crossover implementations without any constraints. More GP operations are needed.