Iterative Relocation For mclust/ mclass

DESCRIPTION:

Performs iterative relocation for a given clustering criterion and classification.

USAGE:

mreloc(classification, x, method = "S*", signif = rep(0, ncol(x)), 
       noise = F, scale = rep(1,ncol(x)), 
       shape = c( 1, rep(.2, (ncol(x)-1))), workspace = 10*nrow(x), 
       iterations = nrow(x)) 

REQUIRED ARGUMENTS:

classification
the output of mclass, or else an integer vector giving the classification for each data point (e.g., the classification component of the output from mclass).
x
n by p matrix containing n p-dimensional data points.

OPTIONAL ARGUMENTS:

method
a character string to select the clustering criterion. The options are the model-based options from mclust: "S*", "S", "spherical" (with varying sizes), "sum of squares" or "trace" (Ward's method), "unconstrained", and "determinant". Only enough of the string to determine a unique match is required. The default is "S*".
signif
vector giving the number of significant decimal places in each component of x. Nonpositive components are allowed. This is used in initializing clustering in some methods.
noise
indicates whether or not Poisson noise should be assumed.
scale
vector for scaling the observations. The ith column of x is multiplied by scale[i] before cluster analysis. The default is rep(1, ncol(x)).
shape
p vector determining the shape of clusters for methods "S*" and "S". The default is c(1, rep( .2, (ncol(x)-1))).
workspace
size of the workspace provided to the underlying Fortran program. The default is 10 times the number of data points.
iterations
desired number of iterations. The default is equal to the number of data points.

VALUE:

an integer vector in which the kth component identifies the new classification of the kth object.

NOTE:

Although all options are allowed, method , noise, error, scale, and shape would usually be expected to be the same as the input to mclust.

SEE ALSO:

, .

EXAMPLES:

years <- c("1960", "1964", "1968", "1972", "1976") 
votes.clust <- mclust(votes.repub[,years], method = "S", noise = T) 
# plot the awe on the current graphics device 
plot(x = 1:length(votes.clust$awe), y = votes.clust$awe) 
votes.class <- mclass(votes.clust, 3) 
votes.reloc <- mreloc(votes.class, votes.repub[,years], method = "S",  
                      noise = T)