Support box constraints?

[tomwenseleers]

Would you see any possibility to support box constraints on the fitted coefficients, similar to what can be achieved with lower.limits and upper.limits in glmnet?

[jaredhuling]

I think it would be possible, but I won't be able to implement it in the near future

[tomwenseleers]

No worries - it was mainly nonnegativity constraints that I needed - I guess for now I'll just have to clip negative coefficients to zero then...

[jaredhuling]

Adding the constraints should be easier than I originally thought; but I still won't be able to get to it for a couple of weeks I think