wolffd@0: function [y, extra, invhess] = netevfwd(w, net, x, t, x_test, invhess)
wolffd@0: %NETEVFWD Generic forward propagation with evidence for network
wolffd@0: %
wolffd@0: %	Description
wolffd@0: %	[Y, EXTRA] = NETEVFWD(W, NET, X, T, X_TEST) takes a network data
wolffd@0: %	structure  NET together with the input X and target T training data
wolffd@0: %	and input test data X_TEST. It returns the normal forward propagation
wolffd@0: %	through the network Y together with a matrix EXTRA which consists of
wolffd@0: %	error bars (variance) for a regression problem or moderated outputs
wolffd@0: %	for a classification problem.
wolffd@0: %
wolffd@0: %	The optional argument (and return value)  INVHESS is the inverse of
wolffd@0: %	the network Hessian computed on the training data inputs and targets.
wolffd@0: %	Passing it in avoids recomputing it, which can be a significant
wolffd@0: %	saving for large training sets.
wolffd@0: %
wolffd@0: %	See also
wolffd@0: %	MLPEVFWD, RBFEVFWD, GLMEVFWD, FEVBAYES
wolffd@0: %
wolffd@0: 
wolffd@0: %	Copyright (c) Ian T Nabney (1996-2001)
wolffd@0: 
wolffd@0: func = [net.type, 'evfwd'];
wolffd@0: net = netunpak(net, w);
wolffd@0: if nargin == 5
wolffd@0:   [y, extra, invhess] = feval(func, net, x, t, x_test);
wolffd@0: else
wolffd@0:   [y, extra, invhess] = feval(func, net, x, t, x_test, invhess);
wolffd@0: end