06 Nov 2002
02 Dec 2002)
Pattern analysis toolbox.
function net = mlpinit(net, prior)
%MLPINIT Initialise the weights in a 2-layer feedforward network.
% NET = MLPINIT(NET, PRIOR) takes a 2-layer feedforward network NET and
% sets the weights and biases by sampling from a Gaussian distribution.
% If PRIOR is a scalar, then all of the parameters (weights and biases)
% are sampled from a single isotropic Gaussian with inverse variance
% equal to PRIOR. If PRIOR is a data structure of the kind generated by
% MLPPRIOR, then the parameters are sampled from multiple Gaussians
% according to their groupings (defined by the INDEX field) with
% corresponding variances (defined by the ALPHA field).
% See also
% MLP, MLPPRIOR, MLPPAK, MLPUNPAK
% Copyright (c) Ian T Nabney (1996-2001)
sig = 1./sqrt(prior.index*prior.alpha);
w = sig'.*randn(1, net.nwts);
elseif size(prior) == [1 1]
w = randn(1, net.nwts).*sqrt(1/prior);
error('prior must be a scalar or a structure');
net = mlpunpak(net, w);