GPR

Trainable mapping for Gaussian Process regression

       W = GPR(A,KERNEL,S_noise)
       W = A*GPR([],KERNEL,S_noise)
       W = A*GPR(KERNEL,S_noise)

 INPUT
   A Dataset used for training
   KERNEL Untrained mapping to compute kernel by A*(A*KERNEL)
            during training, or B*(A*KERNEL) during evaluation with
            dataset B
   S_noise Standard deviation of the noise

Output
W	Mapping: Gaussian Process regression

Description

Fit a Gaussian Process regressor on dataset A. For a nonlinear regressor,  define kernel mapping KERNEL. For kernel definitions, have a look at  proxm.m.

See also

datasets, mappings, svmr, proxm, linearr, testr, plotr,