1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
|
\name{combiPlot}
\alias{combiPlot}
\title{
Plot Classifications Corresponding to Successive Combined Solutions
}
\description{
Plot classifications corresponding to successive combined solutions.
}
\usage{
combiPlot(data, z, combiM, \dots)
}
\arguments{
\item{data}{
The data.
}
\item{z}{
A matrix whose [i,k]th entry is the probability that observation i in the data belongs to the kth class, for the initial solution (ie before any combining). Typically, the one returned by \code{Mclust}/BIC.
}
\item{combiM}{
A "combining matrix" (as provided by \code{\link{clustCombi}}), ie a matrix whose kth row contains only zeros, but in columns corresponding to the labels of the classes in the initial solution to be merged together to get the combined solution.
}
\item{\dots}{
Other arguments to be passed to the \code{\link{Mclust}} plot functions.
}
}
\value{
Plot the classifications obtained by MAP from the matrix \code{t(combiM \%*\% t(z))}, which is the matrix whose [i,k]th entry is the probability that observation i in the data belongs to the kth class, according to the combined solution obtained by merging (according to \code{combiM}) the initial solution described by \code{z}.
}
\references{
J.-P. Baudry, A. E. Raftery, G. Celeux, K. Lo and R. Gottardo (2010). Combining mixture components for clustering. \emph{Journal of Computational and Graphical Statistics, 19(2):332-353.}
}
\author{
J.-P. Baudry, A. E. Raftery, L. Scrucca
}
\seealso{
\code{\link{clustCombi}}, \code{\link{combMat}}, \code{\link{clustCombi}}
}
\examples{
\donttest{
data(Baudry_etal_2010_JCGS_examples)
MclustOutput <- Mclust(ex4.1)
MclustOutput$G # Mclust/BIC selected 6 classes
par(mfrow=c(2,2))
combiM0 <- diag(6) # is the identity matrix
# no merging: plot the initial solution, given by z
combiPlot(ex4.1, MclustOutput$z, combiM0, cex = 3)
title("No combining")
combiM1 <- combMat(6, 1, 2) # let's merge classes labeled 1 and 2
combiM1
combiPlot(ex4.1, MclustOutput$z, combiM1)
title("Combine 1 and 2")
# let's merge classes labeled 1 and 2, and then components labeled (in this
# new 5-classes combined solution) 1 and 2
combiM2 <- combMat(5, 1, 2) \%*\% combMat(6, 1, 2)
combiM2
combiPlot(ex4.1, MclustOutput$z, combiM2)
title("Combine 1, 2 and then 1 and 2 again")
plot(0,0,type="n", xlab = "", ylab = "", axes = FALSE)
legend("center", legend = 1:6,
col = mclust.options("classPlotColors"),
pch = mclust.options("classPlotSymbols"),
title = "Class labels:")}
}
\keyword{cluster}
|
