prtClusterKmeans   Kmeans clustering object
 
     CLUSTER = prtClusterKmeans returns a Kmeans clustering object.
 
     CLUSTER = prtClusterKmeans(PROPERTY1, VALUE1, ...) constructs a
     prtClassFld object CLASSIFIER with properties as specified by
     PROPERTY/VALUE pairs.
 
     A prtClusterKmeans object inherits all properties from the abstract
     class prtCluster. In addition is has the following properties:
 
     nClusters                 - Number of cluster centers to learn 
 
     kmeansHandleEmptyClusters - Speficies operation when degerate clusters found
                                 occur during training.  Allowed values are 'remove'
                                 and 'random'.  'remove' eliminates the
                                 empty cluster. 'random' sets the
                                 cluster mean to a random vector.
 
     For information on the K-Means algorithm, please
     refer to the following URL:
 
     http://en.wikipedia.org/wiki/K-means_clustering
 
     A prtClusterKmeans object inherits the TRAIN, RUN, CROSSVALIDATE and
     KFOLDS methods from prtAction. It also inherits the PLOT method from
     prtCluster.
 
     Invoking the RUN method on a prtClusterKmeans object classifies
     the input data by assigning each observation a label according to
     the cluster center it is closest to. The cluster centers are found
     during training.
 
    Example:
 
    ds = prtDataGenMary                  % Load a prtDataSet
    clusterAlgo = prtClusterKmeans;      % Create a prtClusterKmeans object
    clusterAlgo.nClusters = 3;           % Set the number of desired clusters
 
    % Set the internal decision rule to be MAP. Not required for
    % clustering, but necessary to plot the results.
    clusterAlgo.internalDecider = prtDecisionMap;
    clusterAlgo = clusterAlgo.train(ds); % Train the cluster algorithm
    plot(clusterAlgo);                   % Plot the results