Use the features selected with RFE SVM linear for prediction of SVM rbf

by rugrag   Last Updated October 09, 2018 21:19 PM

I was wondering if the features selected with RFE with SVM linear kernel are still "good" features when we use a non linear model, like SVM rbf kernel. This comes in practice when you want to use SVM as a classifier for the RFE but you are forced to stick to linear kernel: maybe you can do the selection with SVM linear and then the prediction with SVM rbf? If I would answer I would say yes: the features selected with linear SVM are able to explain a linear relationship between descriptors and output, so if we use a non linear model they are still useful. If you can give an explation, it will be quite appreciated.



Related Questions


"...if the data is linearly separable"

Updated June 19, 2015 10:08 AM

Regularized linear vs. RKHS-regression

Updated February 12, 2018 20:19 PM


Using kernels to boost up dimensionality

Updated March 28, 2017 16:19 PM

SVM basic theory?

Updated June 04, 2015 11:08 AM