wolffd@0: wolffd@0:
wolffd@0:wolffd@0: net = gpinit(net, trin, trtargets, prior) wolffd@0: net = gpinit(net, trin, trtargets, prior) wolffd@0:wolffd@0: wolffd@0: wolffd@0:
net = gpinit(net, trin, trtargets)
takes a Gaussian Process data structure net
wolffd@0: together
wolffd@0: with a matrix trin
of training input vectors and a matrix trtargets
of
wolffd@0: training target
wolffd@0: vectors, and stores them in net
. These datasets are required if
wolffd@0: the corresponding inverse covariance matrix is not supplied to gpfwd
.
wolffd@0: This is important if the data structure is saved and then reloaded before
wolffd@0: calling gpfwd
.
wolffd@0: Each row
wolffd@0: of trin
corresponds to one input vector and each row of trtargets
wolffd@0: corresponds to one target vector.
wolffd@0:
wolffd@0: net = gpinit(net, trin, trtargets, prior)
additionally initialises the
wolffd@0: parameters in net
from the prior
data structure which contains the
wolffd@0: mean and variance of the Gaussian distribution which is sampled from.
wolffd@0:
wolffd@0:
x
wolffd@0: and targets t
:
wolffd@0: wolffd@0: wolffd@0: net = gp(2, 'sqexp'); wolffd@0: net = gpinit(net, x, t); wolffd@0: % Train the network wolffd@0: save 'gp.net' net; wolffd@0:wolffd@0: wolffd@0: Another Matlab program can now read in the network and make predictions on a data set wolffd@0:
testin
:
wolffd@0: wolffd@0: wolffd@0: load 'gp.net'; wolffd@0: pred = gpfwd(net, testin); wolffd@0:wolffd@0: wolffd@0: wolffd@0:
gp
, gpfwd
Copyright (c) Ian T Nabney (1996-9) wolffd@0: wolffd@0: wolffd@0: wolffd@0: