wolffd@0: [mask, prior] = rbfprior(rbfunc, nin, nhidden, nout, aw2, ab2)wolffd@0: wolffd@0: wolffd@0:
[mask, prior] = rbfprior(rbfunc, nin, nhidden, nout, aw2, ab2)
wolffd@0: generates a vector
wolffd@0: mask that selects only the output
wolffd@0: layer weights. This is because most uses of RBF networks in a Bayesian
wolffd@0: context have fixed basis functions with the output layer as the only
wolffd@0: adjustable parameters. In particular, the Neuroscale output error function
wolffd@0: is designed to work only with this mask.
wolffd@0:
wolffd@0: The return value
wolffd@0: prior is a data structure,
wolffd@0: with fields prior.alpha and prior.index, which
wolffd@0: specifies a Gaussian prior distribution for the network weights in an
wolffd@0: RBF network. The parameters aw2 and ab2 are all
wolffd@0: scalars and represent the regularization coefficients for two groups
wolffd@0: of parameters in the network corresponding to
wolffd@0: second-layer weights, and second-layer biases
wolffd@0: respectively. Then prior.alpha represents a column vector of
wolffd@0: length 2 containing the parameters, and prior.index is a matrix
wolffd@0: specifying which weights belong in each group. Each column has one
wolffd@0: element for each weight in the matrix, using the standard ordering as
wolffd@0: defined in rbfpak, and each element is 1 or 0 according to
wolffd@0: whether the weight is a member of the corresponding group or not.
wolffd@0:
wolffd@0:
rbf, rbferr, rbfgrad, evidenceCopyright (c) Ian T Nabney (1996-9) wolffd@0: wolffd@0: wolffd@0: wolffd@0: