Automatic radial basis LIBSVM, using NULIBSVC and the Parzen kernel
[W,KERNEL] = PKLIBSVC(A,ALF)
This routine provides a radial basis support vector classifier based on LIBSVC. It estimates the kernel width SIGMA by the the value found by PARZENC. The kernel width used is ALF*3*SQRT(2)*SIGMA. This is much faster than the gridsearch used by RBLIBSVC and may performs equally well.
Note that LIBSVC is a multi-class classifier and thereby uses a single kernel.