Cluster based classification

`semisupc`	This routine trains a user defined classifier on the basis of a labeled training set and a user defined cluster analysis of a large unlabeled dataset.
`clustc`	Classification of objects on the basis of combining multilevel cluster labels and for some objects the true class labels.
`clusthc`	This routine produces a classifier for assigning cluster labels to new objects on the basis of an agglomerative hierarchical clustering result. The classifiers uses the linkage type used in the clustering.
`nnprotoc`	Constructs a 1-nearest neighbor classifier to assign cluster labels to new objects on the basis of the prototypes resulting from a clustering.