prtKernelRbf  Radial basis function kernel
 
   KERNOBJ = prtKernelRbf Generates a kernel object implementing a
   radial basis function.  Kernel objects are widely used in several
   prt classifiers, such as prtClassRvm and prtClassSvm.  RBF kernels
   implement the following function for 1 x N vectors x1 and x2:
 
    k(x1,x2) = exp(-sum((x1-x2).^2)./sigma.^2);
 
   KERNOBJ = prtKernelRbf(PROPERTY1, VALUE1, ...) constructs a
   prtKernelRbfNdimensionScale object KERNOBJ with properties as specified by
   PROPERTY/VALUE pairs. prtKernelRbf objects have the following
   user-settable properties:
 
   KERNOBJ = prtKernelRbf(param,value,...) with parameter value
   strings sets the relevant fields of the prtKernelRbf object to have
   the corresponding values.  prtKernelRbf objects have the following
   user-settable properties:
 
    sigma   - Positive scalar value specifying the width of the
              Gaussian kernel in the RBF function.  (Default value is 1)
 
   Radial basis function kernels are widely used in the machine
   learning literature. For more information on these kernels, please
   refer to:
    
   http://en.wikipedia.org/wiki/Support_vector_machine#Non-linear_classification
 
    prtKernelRbf objects inherit the TRAIN and RUN methods from prtKernel.
 
    % Example
    ds = prtDataGenBimodal;       % Generate a dataset
    k1 = prtKernelRbf;            % Create a prtKernel object with 
                                  % default value of sigma 
    k2 = prtKernelRbf('sigma',2); % Create a prtKernel object with
                                  % specified value of sigma
    
    k1 = k1.train(ds); % Train
    g1 = k1.run(ds); % Evaluate
 
    k2 = k2.train(ds); % Train
    g2 = k2.run(ds); % Evaluate
 
    subplot(2,1,1); imagesc(g1.getObservations);  %Plot the results
    subplot(2,1,2); imagesc(g2.getObservations);