Matthieu Gomez

Don’t hesitate to do pull request at StatsBase. They were pretty receptive when I wanted to add some methods on RegressionModels.

Done (although three state variables is much slower)

To solve your model, I needed to add the constraint that the drift of capital must be non negative at k = 0. This constraint is necessary because, otherwise, at...

The Readme now lists the paper

It's a bit tricky because Inf is meaningful. I will leave this issue open so that people can report if they encounter the same issue.

Yes, it should not be too hard to do — you're welcome to try adding it if you have some time. Here is the code to modify: https://github.com/FixedEffects/FixedEffectModels.jl/blob/eb7bf7318f65841446821c4b7a82820a00b803fc/src/FixedEffectModel.jl#L75

Agreed! However, Metal.jl does not support atomics yet (i.e. the `atomic_add!` function, which is at the core of the algorithm). I will have another look once it does. See https://github.com/JuliaGPU/Metal.jl/issues/79

The package now supports Mac GPUs it but it ends up being slower than CPUs. If anyone is interested in making this faster, the relevant code is https://github.com/FixedEffects/FixedEffects.jl/blob/master/ext/MetalExt.jl

I am not sure there is an issue with the current implementation. The way I think about it is that there is a formula for each Qgram distance (given in...

Sure. I'm also open to change things — following what other libraries typically do in this case. I will leave this issue open.