Prev: Data Concurrency Issues
Next: Scatter plot - making a contour map of point concentration
From: Wangyabf Zhang on 21 Jan 2010 09:30
The svmtrain does not apply the box constraint as in the traditional soft-margin SVM:
norm(a,inf)<=C
instead, according to the code, it seems to replace the box constraint with
(1/2C)*norm(a,2)^2
in the objective function rather than the constraint.

Can anyone give a reference for this implementation of SVM?
 | 
Pages: 1
Prev: Data Concurrency Issues
Next: Scatter plot - making a contour map of point concentration