rdrr1990 / bigKRLS

Now on CRAN, bigKRLS combines bigmemory & RcppArmadillo (C++) for speed into a new Kernel Regularized Least Squares algorithm. Slides:
https://sites.google.com/site/petemohanty/software
28 stars 7 forks source link

Can KRLS be used for classification? #39

Open AmirhosseinModabbernia opened 4 years ago

AmirhosseinModabbernia commented 4 years ago

Hi there,

Can this algorithm be used for classifying a binary variable or is it only suitable for regression of continuous outcomes?

rbshaffer commented 4 years ago

Yes, KRLS can be used with a binary dependent variable (see Section 6 of Hainmueller and Hazlett https://www.jstor.org/stable/pdf/24573219.pdf?casa_token=Exo8zHuL4zcAAAAA:ffH6wLZMNRqPItHsWaHNaH5EtASDZqDOktUXdJ9uU2V41DDKX21FEkOVlGhzreFzPRhUuFJpq8ZAaWHEgZ_5LTD01L20nCCWi6FpeJJ6R92vVq8UYA for an example). However, you may also want to look at KRLS-Logit, which is a work-in-progress implementation of logistic regression in the KRLS framework (see here https://github.com/lukesonnet/KRLS for details).

On Fri, May 22, 2020 at 2:31 PM AmirhosseinModabbernia < notifications@github.com> wrote:

Hi there,

Can this algorithm be used for classifying a binary variable or is it only suitable for regression of continuous outcomes?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rdrr1990/bigKRLS/issues/39, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABNR7ZCZJL45DYIBKCTYDXTRS3AHRANCNFSM4NIBXPSA .

-- Postdoctoral Fellow, Perry World House, University of Pennsylvania Website https://rbshaffer.github.io/

AmirhosseinModabbernia commented 4 years ago

Thanks very much.