weightplots: Ordered posterior plots

In fpc: Flexible Procedures for Clustering

Ordered posterior plots

Description

Ordered posterior plots for Gaussian mixture components, see Hennig (2010).

Usage

  weightplots(z, clusternumbers="all", clustercol=2,
                        allcol=grey(0.2+((1:ncol(z))-1)*
                          0.6/(ncol(z)-1)),
                        lty=rep(1,ncol(z)),clusterlwd=3,
                        legendposition="none",
                        weightcutoff=0.01,ask=TRUE, ...)

Arguments

z	matrix with rows corresponding to observations and columns corresponding to mixture components. Entries are probabilities that an observation has been generated by a mixture component. These will normally be estimated a posteriori probabilities, as generated as component z of the output object from summary.mclustBIC.
clusternumbers	"all" or vector of integers. Numbers of components for which plots are drawn.
clustercol	colour used for the main components for which a plot is drawn.
allcol	colours used for respective other components in plots in which they are not main components.
lty	line types for components.
clusterlwd	numeric. Line width for main component.
legendposition	"none" or vector with two coordinates in the plot, where a legend should be printed.
weightcutoff	numeric between 0 and 1. Observations are only taken into account for which the posterior probability for the main component is larger than this.
ask	logical. If TRUE, it sets par(ask=TRUE) in the beginning and par(ask=FALSE) after all plots were showed.
...	further parameters to be passed on to legend.

Details

Shows posterior probabilities for observations belonging to all mixture components on the y-axis, with points ordered by posterior probability for main component.

Value

Invisible matrix of posterior probabilities z from mclustsummary.

Author(s)

Christian Hennig [e-mail] https://www.unibo.it/sitoweb/christian.hennig/en/

References

Hennig, C. (2010) Methods for merging Gaussian mixture components, Advances in Data Analysis and Classification, 4, 3-34.

Examples

  require(mclust)
  require(MASS)
  data(crabs)
  dc <- crabs[,4:8]
  cm <- mclustBIC(crabs[,4:8],G=9,modelNames="EEE")
  scm <- summary(cm,crabs[,4:8])
  weightplots(scm$z,clusternumbers=1:3,ask=FALSE)
  weightplots(scm$z,clusternumbers=1:3,allcol=1:9, ask=FALSE,
              legendposition=c(5,0.7))
# Remove ask=FALSE to have time to watch the plots.

fpc documentation built on Sept. 24, 2024, 9:07 a.m.