wolffd@0: function [y, a] = glmfwd(net, x)
wolffd@0: %GLMFWD	Forward propagation through generalized linear model.
wolffd@0: %
wolffd@0: %	Description
wolffd@0: %	Y = GLMFWD(NET, X) takes a generalized linear model data structure
wolffd@0: %	NET together with a matrix X of input vectors, and forward propagates
wolffd@0: %	the inputs through the network to generate a matrix Y of output
wolffd@0: %	vectors. Each row of X corresponds to one input vector and each row
wolffd@0: %	of Y corresponds to one output vector.
wolffd@0: %
wolffd@0: %	[Y, A] = GLMFWD(NET, X) also returns a matrix A  giving the summed
wolffd@0: %	inputs to each output unit, where each row corresponds to one
wolffd@0: %	pattern.
wolffd@0: %
wolffd@0: %	See also
wolffd@0: %	GLM, GLMPAK, GLMUNPAK, GLMERR, GLMGRAD
wolffd@0: %
wolffd@0: 
wolffd@0: %	Copyright (c) Ian T Nabney (1996-2001)
wolffd@0: 
wolffd@0: % Check arguments for consistency
wolffd@0: errstring = consist(net, 'glm', x);
wolffd@0: if ~isempty(errstring);
wolffd@0:   error(errstring);
wolffd@0: end
wolffd@0: 
wolffd@0: ndata = size(x, 1);
wolffd@0: 
wolffd@0: a = x*net.w1 + ones(ndata, 1)*net.b1;
wolffd@0: 
wolffd@0: switch net.outfn
wolffd@0: 
wolffd@0:   case 'linear'     % Linear outputs
wolffd@0:     y = a;
wolffd@0: 
wolffd@0:   case 'logistic'   % Logistic outputs
wolffd@0:     % Prevent overflow and underflow: use same bounds as glmerr
wolffd@0:     % Ensure that log(1-y) is computable: need exp(a) > eps
wolffd@0:     maxcut = -log(eps);
wolffd@0:     % Ensure that log(y) is computable
wolffd@0:     mincut = -log(1/realmin - 1);
wolffd@0:     a = min(a, maxcut);
wolffd@0:     a = max(a, mincut);
wolffd@0:     y = 1./(1 + exp(-a));
wolffd@0: 
wolffd@0:   case 'softmax'   	% Softmax outputs
wolffd@0:     nout = size(a,2);
wolffd@0:     % Prevent overflow and underflow: use same bounds as glmerr
wolffd@0:     % Ensure that sum(exp(a), 2) does not overflow
wolffd@0:     maxcut = log(realmax) - log(nout);
wolffd@0:     % Ensure that exp(a) > 0
wolffd@0:     mincut = log(realmin);
wolffd@0:     a = min(a, maxcut);
wolffd@0:     a = max(a, mincut);
wolffd@0:     temp = exp(a);
wolffd@0:     y = temp./(sum(temp, 2)*ones(1,nout));
wolffd@0:     % Ensure that log(y) is computable
wolffd@0:     y(y<realmin) = realmin;
wolffd@0: 
wolffd@0:   otherwise
wolffd@0:     error(['Unknown activation function ', net.outfn]);
wolffd@0: end