DCLUSTK

Dissimilarity based clustering by k-centres or k-medoids

    LAB = DCLUSTK(D,K,TYPE,L)
    LAB = D*DCLUSTK(K,TYPE,L)

Input
D	Square dissimilarity matrix, size M*M
K	Vector of length N with desired numbers of clusters,  default is a sampling of [2:M]
TYPE	'kcentres', k-centres (default)  'kmedoids', k-medoids
L	Vector length max(K), indices of initial centres. Default:  initialisation by DSELPROTO.

Output
LAB	M*N array with the results of the multi-level clusterings for the
M	objects. The columns refer to the N clusterings. They yield for  the objects the prototype indices of the clusters they belong to.

Description

Finds K prototypes from a symmetric distance matrix D. The prototypes  are chosen from all M objects such that the maximum (for k-centres) or  mean (for k-medoids) of the distances over all objects to the nearest  prototype is minimized.

In case K is a vector (length N), the procedure is repeated for as set of  values for K between 2 and M.

See also

datasets, mappings, kmeans, prkmeans, clustk, dclustm, dclusth, dcluste, clusteval, clustcerr, clustc,