Plots to distinguish given classes by ten available projection methods. Includes classical discriminant coordinates, methods to project differences in mean and covariance structure, asymmetric methods (separation of a homogeneous class from a heterogeneous one), local neighborhood-based methods and methods based on robust covariance matrices. One-dimensional data is plotted against the cluster number.
plotcluster(x, clvecd, clnum=NULL,
method=ifelse(is.null(clnum),"dc","awc"),
bw=FALSE,
ignorepoints=FALSE, ignorenum=0, pointsbyclvecd=TRUE,
xlab=NULL, ylab=NULL,
pch=NULL, col=NULL, ...)the data matrix; a numerical object which can be coerced to a matrix.
vector of class numbers which can be coerced into
integers; length must equal
nrow(xd).
one of
usual discriminant coordinates, see discrcoord,
Bhattacharyya coordinates, first coordinate showing
mean differences, second showing covariance matrix differences,
see batcoord,
variance dominated Bhattacharyya coordinates,
see batcoord,
added mean and variance differences optimizing
coordinates, see mvdcoord,
asymmetric discriminant coordinates, see
adcoord,
asymmetric discriminant coordinates with weighted
observations, see awcoord,
asymmetric discriminant coordinates with weighted
observations and robust MCD-covariance matrix,
see awcoord,
neighborhood based coordinates,
see ncoord,
neighborhood based coordinates with weighted neighborhoods,
see ncoord,
asymmetric neighborhood based coordinates,
see ancoord.
Note that "bc", "vbc", "adc", "awc", "arc" and "anc" assume that there are only two classes.
integer. Number of the class which is attempted to plot
homogeneously by "asymmetric methods", which are the methods
assuming that there are only two classes, as indicated above.
clnum is ignored for methods "dc" and "nc".
logical. If TRUE, the classes are distinguished by
symbols, and the default color is black/white.
If FALSE, the classes are distinguished by
colors, and the default symbol is pch=1.
logical. If TRUE, points with label
ignorenum in clvecd are ignored in the computation for
method and are only projected afterwards onto the resulting
units. If pch=NULL, the plot symbol for these points is "N".
one of the potential values of the components of
clvecd. Only has effect if ignorepoints=TRUE, see above.
logical. If TRUE and pch=NULL
and/or col=NULL, some hopefully suitable
plot symbols (numbers and letters) and colors are chosen to
distinguish the values of clvecd, starting with "1"/"black"
for the cluster with the smallest clvecd-code (note that
colors for clusters with numbers larger than minimum number
+3 are drawn at random from all available colors).
FALSE produces
potentially less reasonable (but nonrandom) standard colors and symbols if
method is "dc" or "nc", and will only distinguish whether
clvecd=clnum or not for the other methods.
label for x-axis. If NULL, a default text is used.
label for y-axis. If NULL, a default text is used.
plotting symbol, see par.
If NULL, the default is used.
plotting color, see par.
If NULL, the default is used.
additional parameters passed to plot or the
projection methods.
Hennig, C. (2004) Asymmetric linear dimension reduction for classification. Journal of Computational and Graphical Statistics 13, 930-945 .
Hennig, C. (2005) A method for visual cluster validation. In: Weihs, C. and Gaul, W. (eds.): Classification - The Ubiquitous Challenge. Springer, Heidelberg 2005, 153-160.
Seber, G. A. F. (1984). Multivariate Observations. New York: Wiley.
Fukunaga (1990). Introduction to Statistical Pattern Recognition (2nd ed.). Boston: Academic Press.
discrcoord, batcoord,
mvdcoord, adcoord,
awcoord, ncoord,
ancoord.
discrproj is an interface to all these projection methods.
rFace for generation of the example data used below.
# NOT RUN {
set.seed(4634)
face <- rFace(300,dMoNo=2,dNoEy=0)
grface <- as.integer(attr(face,"grouping"))
plotcluster(face,grface)
plotcluster(face,grface==1)
plotcluster(face,grface, clnum=1, method="vbc")
# }
Run the code above in your browser using DataLab