Prev: Efficient reading of binary complex I/Q values
Next: mexing on Mac OS X with g95/gfortran
From: Nicole on 8 Jun 2010 04:49
Hi all,

Using svmtrain/svmclassify I'm returned the class labels. I additionally need the distances from the SVM hyperplane to derive a confidence value for the classification from that, ...

So, is there any way to get the distances from the SVM hyperplane?

An example how to do this would be great - like e.g. an extension of the fisheries example.

Any help will be greatly appreciated!
Thanks,
Nici



(I'm currently trying to re-implement this paper: http://www.cs.tau.ac.il/~wolf/papers/stereoconversion_web.pdf)
From: Nicole on 10 Jun 2010 08:52
incase anybody needs this as well, see svmdecision.m
http://www.google.com/codesearch/p?hl=de#tZmoENZ8kcc/trunk/svmdecision.m&q=svmdecision%20lang:matlab&sa=N&cd=1&ct=rc
 | 
Pages: 1
Prev: Efficient reading of binary complex I/Q values
Next: mexing on Mac OS X with g95/gfortran