[FEA] Add support for LedoitWolf

[will-hill]

In relation to FSI, specifically Portfolio Optimization, implementing the sklearn.covariance.* packages would be helpful for shrinkage on covariance matrices. Starting with LedoitWolf would be helpful since it is a modern algorithm for shrinkage.

Here is the Scikit-Learn class impl:
https://scikit-learn.org/stable/modules/generated/sklearn.covariance.LedoitWolf.html#sklearn.covariance.LedoitWolf

I tested another open-source impl simply swapping numpy with cupy and the speedup was significant.
https://github.com/WLM1ke/LedoitWolf

This version differs in default parameters from the Scikit-Learn version but close enough to showcase CUDA-X acceleration.

Link to publication:
http://www.ledoit.net/honey.pdf

[csadorf]

Thank you for the suggestion and the links to the reference implementations! We will take it into consideration.