Implementation of sklearn.linear_model.ElasticNet

[PaulJWright]

Describe the feature

One of the most simple examples could be to implement an interface with sklearn.linear_model.ElasticNet: https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.ElasticNet.html as used in https://iopscience.iop.org/article/10.3847/1538-4357/ad1837#apjad1837f5

Proposed solution

No response

[advait-zx]

i can work on it

[nabobalis]

i can work on it

Hello @advait-zx, you are welcome to try this but I think an issue like this requires some knowledge of sunkit-dem but also the use case for this library. So it might be a tricky issue for someone who is new.

[advait-zx]

well, I can give it a try

[nabobalis]

Sure. Good luck!

I can try to help but this package is pretty left on its own these days. Not many people are aware of how or if it even works.

[advait-zx]

what kind of Proposed solution is needed and also mention specific requirment

[nabobalis]

That is a great question.

I think first you'll need to figure how on earth someone uses this library.
Then create a python example of that.
From there, see how methods are chosen by users.
From that, see how the methods are coded in the library.
Then, add a new method that uses ElasticNet underneath.

Does that help in anyway?

[PaulJWright]

I didn't have a proposed solution for this. I would first look at the paper and understand what they're doing.

There are a few issues that we need to resolve for this package in general. I have a working understanding of the implementation by @wtbarnes, but this package needs dedicated dev time. Once a single DEM method is implemented, the others should be straight forward as it utilises the factory method. This was an initial idea of a simple-to-implement approach that doesn't require coding up the minimisation problem

[alasdairwilson]

There isn't really an existing implementation you can follow in the package which is really too early of a stage to get to grips with.

For some reason I don't have write to this repo but this issue should probably have been tagged with expert package knowledge and solar physics knowledge.

You are obviously still welcome to go for it if you are confident.

[PaulJWright]

There isn't really an existing implementation you can follow in the package which is really too early of a stage to get to grips with.

For some reason I don't have write to this repo but this issue should probably have been tagged with expert package knowledge and solar physics knowledge.

You are obviously still welcome to go for it if you are confident.

Can you tag this, I don't have permission