You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When working with random Fourier features (RFs) for the approximation of a Gaussian kernel, it would be very useful to have the following functionalities:
A field of opt to set the sigma parameter used in the random feature mapping.
A paramsel method such as paramsel_siglamho, but working for RFs.
The text was updated successfully, but these errors were encountered:
raffaello-camoriano
changed the title
Random features for Gaussian kernel approximation functions provide no sigma parameter exmplicit choice, nor associated paramsel method
Random features for Gaussian kernel approximation functions provide no sigma parameter explicit choice, nor associated paramsel method
Jun 24, 2015
When working with random Fourier features (RFs) for the approximation of a Gaussian kernel, it would be very useful to have the following functionalities:
The text was updated successfully, but these errors were encountered: