-
Notifications
You must be signed in to change notification settings - Fork 28
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FEAT - Add GroupLasso and GroupLogisticRegression estimators #197
Comments
Hello, following up on the issue from Celer, it would be great to be able to run GroupLasso with non-neg weights. The standard Lasso already uses For my application I also need to run on sparse X input matrices. From what I have seen, some of the regressors can handle the |
Hello,
|
Thank you, I am glad that it is doable. Is there anything I can do to help? |
@tomaszkacprzak we can start by supporting sparse matrices for the GroupBCD solver. Do you think you could send a PR doing that ? You'd need to implement See current implementations for CD (not BCD) in Then inside the solver we just have an if when doing the epoch of coordinate descent : https://github.com/scikit-learn-contrib/skglm/blob/main/skglm/solvers/anderson_cd.py#L147 Do not hesitate to start a very WIP PR, so that we can discuss over there! |
In the meantime, I will open a separate PR for the positive weights |
Do we want a |
I think it's a popular enough model such that the visibility benefits of implementing its class overweighs the cost of maintenance: I am +1 on the |
@tomaszkacprzak did you try our implementation of |
@mathurinm I just saw it, going to give it a try. Then I'll start the PR for sparse X. Thanks! |
@mathurinm I noticed that in Example for random seed 12323
Any suggestions what can cause this? |
Thanks for the report, this seems to show that the |
This should be fixed now @tomaszkacprzak, thanks @Badr-MOUFAD |
Great, I just checked it out, works well. Thanks @Badr-MOUFAD. I'll start a PR for sparse X. |
No description provided.
The text was updated successfully, but these errors were encountered: