The HPSVM Procedure
KERNEL Statement
KERNEL kernel-type / kernel-parameters;
The KERNEL statement specifies the type of kernel and any associated parameters to be used during training. You can specify one of the following kernel-types; for some kernel-types, you must specify one or two kernel-parameters.
- LINEAR
-
uses a linear kernel during training. No kernel-parameter is needed. The kernel is defined as
where
and 
are two vectors and 
is the inner product. - POLYNOM
-
uses a polynomial kernel during training. Specify the polynomial degree as the kernel-parameter. For example, specify KERNEL POLYNOM / DEGREE=2. The kernel is defined as
where p is the degree of the polynomial.
- RBF
-
uses a radial basis function (RBF) kernel during training. Specify K_PAR as the kernel-parameter. For example, specify KERNEL RBF / K_PAR=1. The kernel is defined as
where
is a positive number, which is K_PAR. - SIGMOID
-
uses a sigmoid kernel during training. Specify two kernel-parameters, K_PAR and K_PAR2. For example, specify KERNEL SIGMOID / K_PAR=1 K_PAR2=–1. The kernel is defined as
where
is a positive number(as K_PAR) and 
is a negative number(as K_PAR2).
The default kernel type is LINEAR.