Giters

jasonge27 / picasso

Penalized Sparse Learning Solver - Unleash the Power of Nonconvex Penalty

Home Page:https://jasonge27.github.io/picasso/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Support box or nonnegativity constraints?

tomwenseleers opened this issue · comments

commented

Would it be possible for picasso in the future to also support box constraints on the fitted coefficients by any chance, similar to what glmnet can do using arguments lower.limits and upper.limits? E.g. to support nonnegativity constraints?

commented

A staightforward implementation is to add the clipping mechanism in the coordinate_descent function implemented in objective function classes https://github.com/jasonge27/picasso/blob/f4b6a10a5f3e1dbc6797a636d425658563450a9b/src/objective/glm.cpp#L24-L48. I am not sure if clipping works well in terms of stability and accuracy of the algorithm. We will test this implementation in the future. Pull request is highly welcome!