about tuning

[JeffreySarnoff]

Thank you for this package. Robust is helpful :). If you have the information on hand, please give an example or two of tuning for rpca_fista and rpca_admm effectively + efficiently. A sentence or two on when to use which would be helpful to the non-experts.

[baggepinnen]

Unfortunately, I've found them hard to tune and haven't made much use of them myself, hence the limited documentation. The default algorithm "Augmented Lagrange Multiplier Method " exposed through the rpca function, comes with nice theory about the tuning and is, in my experience, much easier to use.

[JeffreySarnoff]

Thank you for the clarification. I will stick with what works, using the version with what you have more experience.

On Tue, Nov 29, 2022, 2:41 AM Fredrik Bagge Carlson < @.***> wrote:

Unfortunately, I've found them hard to tune and haven't made much use of them myself, hence the limited documentation. The default algorithm "Augmented Lagrange Multiplier Method " exposed through the rpca function, comes with nice theory about the tuning and is, in my experience, much easier to use.

— Reply to this email directly, view it on GitHub https://github.com/baggepinnen/RobustFactorizations.jl/issues/2#issuecomment-1330211498, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAM2VRQSEID56KLUVN25HJ3WKWXUJANCNFSM6AAAAAASMSVYR4 . You are receiving this because you authored the thread.Message ID: @.***>