lyngby_km_error

 lyngby_km_error      - K-means, error 

        function [e, mys, Sigmas, ps] = lyngby_km_error(X, a, ...
           'PropertyName', 'PropertyValue')

       Input:  X   Datamatrix, size: examples x variables. ie. if
                   voxels are to be clustered the datamatrix should
                   be voxels x time.
               a   Assignment vector
                Property:
                  'MeanOrigin'     [ {MeanTrainingSet} | ClusterCenters ]
                                   Specifices who the center of the
                                   gaussians should be found.
                  'ClusterCenters' Centers for the clusters. This
                                   should be defined if 'MeanOrigin'
                                   is set to 'ClusterCenters'
                  'VarianceType'   [ {Anisotrop} | Isotrop ]
                  'Reg'            {0} Regularization
                  'TrainingSet'    Indices for the training set. The
                                   first half of the dataset will be
                                   used if it is not defined.
                  'ValSet'         Indices for the validation set

         Output:        e   Error, negative log of p(x)

       See also: lyngby_gmm_fit, lyngby_gmm_error