RBSVC

Trainable automatic radial basis Support Vector Classifier

    [W,KERNEL,NU,C] = RBSVC(A)
    [W,KERNEL,NU,C] = A*RBSVC

Input
A	Dataset

Output
W	Mapping: Radial Basis Support Vector Classifier
KERNEL	Untrained mapping, representing the optimised kernel
NU	Resulting value for NU from NUSVC (W = NUSVC(A,KERNEL,C)
C	Resulting value for C (W = SVC(A,KERNEL,C)

Description

This routine computes a classifier by NUSVC using a radial basis kernel  with an optimised standard deviation by REGOPTC. The resulting classifier  W is identical to NUSVC(A,KERNEL,NU). As the kernel optimisation is based  on internal cross-validation the dataset A should be sufficiently large.  Moreover it is very time-consuming as the kernel optimisation needs  about 100 calls to SVC.

If any class in A has less than 20 objects, the kernel is not optimised  by a grid search but by PKSVM, using the Parzen kernel.

Note that SVC is basically a two-class classifier. The kernel may  thereby be different for all base classifiers and is separately optimised  for each of them.

See also

mappings, datasets, proxm, svc, nusvc, regoptc, pksvm,