Skip to content
An R implementation of the (multiple) Support Vector Machine Recursive Feature Elimination (mSVM-RFE) feature ranking algorithm
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
demo
sge
LICENSE.txt
README.mdown
msvmRFE.R

README.mdown

(multiple) Support Vector Machine Recursive Feature Elimination (mSVM-RFE)

Full documentation is available at: http://www.colbyimaging.com/wiki/statistics/msvm-rfe

This package contains an R implementation of the mSVM-RFE algorithm (Duan et al., 2005), including the option to cut the features by half each round (instead of one-by-one) if there are many features.

The main function is adapted from http://www.uccor.edu.ar/paginas/seminarios/Software/SVM_RFE_R_implementation.pdf.

Also included are tools for wrapping the feature ranking/selection process in an external layer of cross-validation for obtaining unbiased estimates of generalization error/accuracy (See Ambroise et al., 2002).

This entire process can be massively parallelized, and an example implementation is included that uses the SGE cluster interface.

Contact: John Colby

You can’t perform that action at this time.