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))

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.

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.

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)