MATLAB Function Reference | Search  Help Desk |
mkmeans | See Also |
performs multi prototypes clustering
Syntax
[cg,cp] = mkmeans(d,n,p)
[cg,cp] = mkmeans(d,n,p,l)
Description
[cg,cp] = mkmeans(d,n,p)
Finds the center of the clusters cg
and there associated
prototypes cp
using a new clustering method developped in IRIDIA.
d
is the data set, n
the number of clusters and p
the number of prototypes by cluster.
cg
is a n
rows matrix (one for each cluster).
cp
is a p
rows and n
columns 3D matrix.
[cg,cp] = mkmeans(d,n,p,l)
l
is the weight of the center of the cluster in
the definition of the cost function minimised by the clustering method.
kmeans, fkmeans, gkclus