Hi all,

I'm wondering whether anyone has succeeded in building a R-SVM model using a distance or similarity matrix as input rather than descriptors or fingerprints?

Cheers,

Pierre