Open roya0045 opened 7 years ago
Hi,
Thanks for your suggestion. Well, I will be glad if my algorithms can be a part of scipy. But I'am afraid that they are not suitable. My methods are specified at solving a special class of problems. But it seems to me that scipy only contains general-purpose solvers.
Do you think there is a way to adapt some of the code to generalize the solver? I'm converting to python an algorithm that solve the same problems that yours is made to solve and I'm trying to find a way to generalize it too.
I'll keep looking for a way to do something and if I find a way to generalize it I'll update you.
Do you know there is already a Python library (lighting)(http://contrib.scikit-learn.org/lightning/) which solves the same problems?
No I was not aware, though I am not currently working on this type of solver, but I guess it would be fairly easy to implement variable to use Barzilai-Borwein step size instead of the normal one.
Thanks for the heads-up!
Greetings,
I'm just wondering if you have thought about submitting the two algorithms as optimization methods for scipy? Those would make a great addition.