batchkmeans.default: K-means Clustering for vector data
In yasomi: Yet Another Self Organising Map Implementation

Description Usage Arguments Details Value Author(s) See Also Examples

Performs K-means clustering on vector data.

1
2
3

## Default S3 method:
batchkmeans(data, ncenters, init = c("prototypes", "random", "cluster"),
prototypes, weights, max.iter, verbose = FALSE, keepdata = TRUE, ...)

`data`	the data to cluster. This can be, a matrix or data frame of observations
`ncenters`	the number of clusters
`init`	the initialisation method ((see `batchkmeans` for details)
`prototypes`	Initial values for the prototypes (it must have the same number of columns as `data` and its row number must be equal to `ncenters`). If missing, initial prototypes are chosen via the method specified by the `init` parameter (see details)
`weights`	optional weights for the data points
`max.iter`	maximal number of iterations of the algorithm
`verbose`	switch for tracing the clustering process
`keepdata`	if `TRUE`, the original data are returned as part of the result object
`...`	not used

This methods implements the standard (batch) K-means clustering algorithm (more precisely the Lloyd-Forgy algorithm).

An object of class "batchkmeans" and of class "batchkmeansnum", a list with components including

`prototypes`	a matrix containing the coordinates of the prototypes
`classif`	a vector of integer indicating to which cluster each observation has been assigned
`errors`	a vector containing the evolution of the quantisation error during the fitting process
`data`	the original data if the function is called with `keepdata = TRUE`
`weights`	the weights of the data points if the function is called with `keepdata = TRUE` and if the `weights` is given

Fabrice Rossi

See batchsom for Self-Oganising Map which provides both clustering and visualisation and kmeans for the version provided by the stats package.

## inspired by the kmeans example in the stats package
require(graphics)

## 2 dimensional dataset
X <- cbind(rnorm(200,mean=2,sd=0.35),rnorm(200,mean=-1,sd=0.35))
Y <- cbind(runif(200,min=-1.5,max=-0.75),runif(200,min=0,max=0.5))
Z <- cbind(rnorm(200,sd=0.15),rnorm(200,sd=0.5))
M <- matrix(c(sin(pi/4),cos(pi/4),-cos(pi/4),sin(pi/4)),ncol=2)
U <- rbind(X,Y,Z%*%M+c(rep(0.25,200),rep(-0.5,200)))
U <- scale(U)

km <- batchkmeans(U,3)

plot(U,col=km$classif,cex=0.5)
points(km$prototypes,col=1:3,pch=20,cex=2)