Art

Using FLANN requires manual memory management, because it maintains in-memory index. How does it fit into a proposed workflow of creating a model and using it for multiple predictions? It...

Most of the test use real vectors for describing individuals. A binary input is used in [knapsack](https://github.com/wildart/Evolutionary.jl/blob/cf3f2fbf5e1be13d6f5db23eabe5dbfab4d5b688/test/knapsack.jl) test. [This example](https://github.com/wildart/Evolutionary.jl/blob/master/examples/MLP.ipynb) uses a population of neural nets. Yes, you can define...

More like ``` julia @checked_lib MBED_TLS joinpath(dirname(@__FILE__),"usr\\bin64\\libmbedtls.DLL") ```

We can just specify how do we reference dependencies in `build.jl` script by absolute path (in case of HB or WinRPM) or relative.

> That would add complication for every package author who wants to use BinDeps. As if using BinDeps is a walk in park, don't event start.

How do you define your initial population? From the error it looks like your population is defined as matrix, you may want to convert it into vector of vectors.

It's a typo in the docs. Use `MetricMDS` type as a first parameter instead of `MDS`, e.g. ``` mds = fit(MetricMDS, D; distances=true, metric=isotonic) ```

What package version are you using?

I implemented PAV algorithm from Best, M.J., Chakravarti, N. Active set algorithms for isotonic regression; A unifying framework. Mathematical Programming 47, 425–439 (1990). As I recall, sk-learn implementation outputs response...

I see. That was a problem with previous implementation. The labels and indices were conflated which caused bounds errors if labels weren't properly defined, #187. It looks like the design...