SVM Result Meaning in R

Question

I'm new to using the e1071 package for SVM classification in R. I am getting the results below.

Parameters:
     SVM-Type: nu-classification 
   SVM-Kernel: sigmoid 
        gamma: 0.03225806 
       coef.0: 0 
           nu: 0.5 

Number of Support Vectors:  229

Parameters:
   SVM-Type:  nu-classification 
 SVM-Kernel:  polynomial 
     degree:  3 
      gamma:  0.03225806 
     coef.0:  0 
         nu:  0.5 

Number of Support Vectors:  235

what does gamma, coef.0, nu and degree means?

Answer 1

• Gamma and coef.0 are parameters in the kernels.

A sigmoid kernel is tanh(gamma*u'*v + coef0).

Degree is used with a polynomial kernel and indicates the degree of the polynomial. A polynomial kernel is (gamma*u'*v + coef0)^degree.

• nu is a parameter needed for nu-classification.