tclust (version 1.4-1)

plot.ctlcurves: plot Method for ctlcurves Objects

Description

The plot method for class ctlcurves: This function plots a ctlcurves object, comparing the target functions values with different values of parameter restr.fact.

Usage

# S3 method for ctlcurves
plot (x, what = c ("obj", "min.weights", "doubtful"),
         main, xlab, ylab, xlim, ylim, col, lty = 1, ...)

Arguments

x

The ctlcurves object to be printed.

what

A string indicating which type of plot shall be drawn. See the details section for more information.

main

A character-string containing the title of the plot.

xlab, ylab, xlim, ylim

Arguments passed to plot.

col

A single value or vector of line colors passed to lines.

lty

A single value or vector of line types passed to lines.

Arguments to be passed to or from other methods.

Details

These curves show the values of the trimmed classification (log-)likelihoods when altering the trimming proportion alpha and the number of clusters k. The careful examination of these curves provides valuable information for choosing these parameters in a clustering problem. For instance, an appropriate k to be chosen is one that we do not observe a clear increase in the trimmed classification likelihood curve for k with respect to the k+1 curve for almost all the range of alpha values. Moreover, an appropriate choice of parameter alpha may be derived by determining where an initial fast increase of the trimmed classification likelihood curve stops for the final chosen k. A more detailed explanation can be found in Garc<ed>a-Escudero et al. (2010).

This function implements a series of plots, which display characteristic values of the each model, computed with different values for k and alpha. The plot type is selected by setting argument what to one of the following values:

"obj"

Objective function values.

% \item{\code{"out"}}{The number of outlying observations identified among the % clusters. % Here observations are counted, which were assigned to a % cluster, and within this cluster yield a squared % Mahalanobis distance larger than % \code{qchisq (1-mah.alpha, p)}, with % \code{p} the number of dimensions of the data matrix \code{x}, % and \code{mah.alpha} the ?? level as chosen in % function \code{\link{ctlcurves}}.\cr % A dashed line represents the chosen ?? level. % }
"min.weights"

The minimum cluster weight found for each computed model. This plot is intended to spot spurious clusters, which in general yield quite small weights.

"doubtful"

The number of "doubtful" decisions identified by DiscrFact.

References

Garc<ed>a-Escudero, L.A.; Gordaliza, A.; Matr<e1>n, C. and Mayo-Iscar, A. (2010), "Exploring the number of groups in robust model-based clustering." Statistics and Computing, (Forthcoming). Preprint available at www.eio.uva.es/infor/personas/langel.html.

Examples

Run this code
# NOT RUN {
sig <- diag (2)
cen <- rep (1, 2)
x <- rbind(mvtnorm::rmvnorm(108, cen * 0,   sig),
	       mvtnorm::rmvnorm(162, cen * 5,   sig * 6 - 2),
	       mvtnorm::rmvnorm(30, cen * 2.5, sig * 50)
)

ctl <- ctlcurves(x, k = 1:4)
plot(ctl)

# }

Run the code above in your browser using DataCamp Workspace