tclust (version 2.0-5)

plot.DiscrFact: The plot method for objects of class DiscrFact

Description

The plot method for class DiscrFact: Next to a plot of the tclust object which has been used for creating the DiscrFact object, a silhouette plot indicates the presence of groups with a large amount of doubtfully assigned observations. A third plot similar to the standard tclust plot serves to highlight the identified doubtful observations.

Usage

# S3 method for DiscrFact
plot(
  x,
  enum.plots = FALSE,
  xlab = "Discriminant Factor",
  ylab = "Clusters",
  print.DiscrFact = TRUE,
  xlim,
  col.nodoubt = grey(0.8),
  by.cluster = FALSE,
  ...
)

Arguments

x

An object of class DiscrFact as returned from DiscrFact()

enum.plots

A logical value indicating whether the plots shall be enumerated in their title ("(a)", "(b)", "(c)").

xlab, ylab, xlim

Arguments passed to funcion plot.tclust()

print.DiscrFact

A logical value indicating whether each clusters mean discriminant factor shall be plotted

col.nodoubt

Color of all observations not considered as to be assigned doubtfully.

by.cluster

Logical value indicating whether optional parameters pch and col (if present) refer to observations (FALSE) or clusters (TRUE)

...

Arguments to be passed to or from other methods

Details

plot_DiscrFact_p2 displays a silhouette plot based on the discriminant factors of the observations. A solution with many large discriminant factors is not reliable. Such clusters can be identified with this silhouette plot. Thus plot_DiscrFact_p3 displays the dataset, highlighting observations with discriminant factors greater than the given threshold. The function plot.DiscrFact() combines the standard plot of a tclust object, and the two plots introduced here.

References

García-Escudero, L.A.; Gordaliza, A.; Matrán, C. and Mayo-Iscar, A. (2011), "Exploring the number of groups in robust model-based clustering." Statistics and Computing, 21 pp. 585-599, <doi:10.1007/s11222-010-9194-z>

Examples

 sig <- diag (2)
 cen <- rep (1, 2)
 x <- rbind(MASS::mvrnorm(360, cen * 0,   sig),
 	       MASS::mvrnorm(540, cen * 5,   sig * 6 - 2),
 	       MASS::mvrnorm(100, cen * 2.5, sig * 50))

 clus.1 <- tclust(x, k = 2, alpha=0.1, restr.fact=12)
 clus.2 <- tclust(x, k = 3, alpha=0.1, restr.fact=1)

 dsc.1 <- DiscrFact(clus.1)
 plot(dsc.1)

 dsc.2 <- DiscrFact(clus.2)
 plot(dsc.2)