Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: xclust
See also: kmeans adaptive agglom

Quantlet: divisive
Description: performs an adaptive divisive K-means cluster analysis with appropriate (adaptive) multivariate graphic using principal components

Usage: cd = divisive (x, k, w, m, sv)
Input:
x n x p matrix of n points to be clustered
k scalar, number of clusters
w p x 1, matrix of weights for coordinates
m n x 1, matrix of weights (masses) of points
sv scalar, seed value for random numbers
Output:
cd.p n x 1 matrix, partition of n points given in x into k clusters
cd.n k x 1 matrix of number of observations of clusters
cd.a p x 1 matrix of final (pooled) adaptive weights of the variables

Example:
; load the library xclust
library("xclust")
; initialize random generator
randomize(0)
; generate basis data
x  = normal(30, 5)
; generate 4 clusters
x1 = x - #(2,1,3,0,0)'
x2 = x + #(1,1,3,1,0.5)'
x3 = x + #(0,0,1,5,1)'
x4 = x - #(0,2,1,3,0)'
x  = x1|x2|x3|x4
; compute column variances
w  = 1./var(x)
; compute row weights
m  = matrix(rows(x))
; apply divisive
cd = divisive(x, 4, w, m, 1111)
; compare estimated and true partition
conting(cd.p, 1+ceil((1:120)/30))

Result:
Content of object h

[1,]        0        0       30        0        0
[2,]        0        0        0       30        0
[3,]        0       30        0        0        0
[4,]        0        0        0        0       30



Author: H.-J. Mucha, S. Klinke, 20010912
(C) MD*TECH Method and Data Technologies, 05.02.2006