search

mathurinm / celer

Fast solver for L1-type problems: Lasso, sparse Logisitic regression, Group Lasso, weighted Lasso, Multitask Lasso, etc.
https://mathurinm.github.io/celer/
BSD 3-Clause "New" or "Revised" License
199 stars 33 forks source link

GroupLasso with positive=True #290

Closed tomaszkacprzak closed 7 months ago

tomaszkacprzak commented 9 months ago

It would be great to implement GroupLasso with a constraint for positive coefficients. Standard Lasso already implements it. Would that be possible?

mathurinm commented 9 months ago

Hi @tomaszkacprzak, i think it'd be doable, by adding positive to dnorm_grp, set_prios_grp and clipping to 0 when updating w[j] L417 of group_fast.pyx.

However this may be easier to implement in our new package, https://github.com/scikit-learn-contrib/skglm. Have a look at the package and if its usage fits your needs, consider opening a feature request over there; @Badr-MOUFAD maybe has time to have a look !

mathurinm commented 7 months ago

Implemented in skglm.