Daniel@0: function e = mdnerr(net, x, t)
Daniel@0: %MDNERR	Evaluate error function for Mixture Density Network.
Daniel@0: %
Daniel@0: %	Description
Daniel@0: %	 E = MDNERR(NET, X, T) takes a mixture density network data structure
Daniel@0: %	NET, a matrix X of input vectors and a matrix T of target vectors,
Daniel@0: %	and evaluates the error function E. The error function is the
Daniel@0: %	negative log likelihood of the target data under the conditional
Daniel@0: %	density given by the mixture model parameterised by the MLP.  Each
Daniel@0: %	row of X corresponds to one input vector and each row of T
Daniel@0: %	corresponds to one target vector.
Daniel@0: %
Daniel@0: %	See also
Daniel@0: %	MDN, MDNFWD, MDNGRAD
Daniel@0: %
Daniel@0: 
Daniel@0: %	Copyright (c) Ian T Nabney (1996-2001)
Daniel@0: %	David J Evans (1998)
Daniel@0: 
Daniel@0: % Check arguments for consistency
Daniel@0: errstring = consist(net, 'mdn', x, t);
Daniel@0: if ~isempty(errstring)
Daniel@0:   error(errstring);
Daniel@0: end
Daniel@0: 
Daniel@0: % Get the output mixture models
Daniel@0: mixparams = mdnfwd(net, x);
Daniel@0: 
Daniel@0: % Compute the probabilities of mixtures
Daniel@0: probs     = mdnprob(mixparams, t);
Daniel@0: % Compute the error
Daniel@0: e       = sum( -log(max(eps, sum(probs, 2))));
Daniel@0: