MEANSHIFT

Mean shift mode-seeking clustering

    LAB = MEANSHIFT(A,K,NEST)
    LAB = A*MEANSHIFT(K,NEST)

Input
A	Dataset of M objects (rows).
K	Scalar or a vector of length N with neighbor numbers. Default is  a geometric series between 1 and M.
NEST	Logical, if TRUE (default) the output set of clusterings (columns  of LAB) is made nested by RECLUSTN.

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

This is a wrapper around the MeanShiftCluster routine by Bryan Feldman  and Bart Finkston found in MathWork File Exchange.  The width parameter is estimated by the average K-NN distance. K can be a  set of values, resulting in a multilevel clustering, stored in columns of  LAB. Each index in LAB points the object of A that is most close to the  corresponding mode of the cluster found by MeanShiftCluster.

The multilevel clustering can be made nested by RECLUSTN.

See also

mappings, datasets, kmeans, hclust, modeclust, modeclustf, dclustm, reclustn,