In Pairwise Clustering, clustering problems are casted into binary clustering problems of sample pairs. Currently We use "Clustering by Class-Prior Matching (CCPM)" for the binary clusterer, but any clustering algorithms can be employed. One of the advantages of pairwise clustering is that we don't have to determine the number of clusters in advance.