PRTools contents

PRTools manual

svmr

SVMR

SVM regression

       W = SVMR(X,NU,KTYPE,KPAR,EP)

Input
 X Regression dataset
 NU Fraction of objects outside the 'data tube'
 KTYPE Kernel type (default KTYPE='p', for polynomial)
 KPAR Extra parameter for the kernel
 EP Epsilon, with of the 'data tube'

Output
 W Support vector regression

Description

Train an nu-Support Vector Regression on dataset X with parameter NU.  The kernel is defined by kernel type KTYPE and kernel parameter KPAR.  For the definitions of these kernels, have a look at proxm.m.

See also

linearr, proxm, gpr, svc,

PRTools contents

PRTools manual