ks (version 1.4.2)

plot.dade: Density estimates and partition plot for discriminant analysis for 2- and 3-dimensional data

Description

Density estimate and partition classes plot for 2- and 3-dimensional for kernel, linear and quadratic discriminant analysis

Usage

## bivariate
## S3 method for class 'dade':
plot(x, y, y.group, prior.prob=NULL, display="part",
    cont=c(25,50,75), ncont=NULL, xlim, ylim, xlabs="x",
    ylabs="y", drawlabels=TRUE, cex=1, pch, lty, col,
    lcol, ptcol, ...)
## trivariate
## S3 method for class 'dade':
plot(x, y, y.group, prior.prob=NULL, display="rgl",
    cont=c(25,50), colors, alphavec, origin=c(0,0,0), endpts,
    xlabs="x", ylabs="y", zlabs="z", drawpoints=TRUE, size=3,
    ptcol, ...)

Value

• Plot of 2-d density estimates (and partition) for discriminant analysis is sent to graphics window. Plot for 3-d is generated by the misc3d and rgl libraries and is sent to RGL window.

synopsis

plot.dade(x, y, y.group, prior.prob=NULL, display="part", cont=NULL, ncont=NULL, ...)

Details

If prior.prob is set to a particular value then this is used. The default is NULL which means that the sample proportions are used.

The object x contains the training data and its group labels. If y and y.group are missing then the training data points are plotted. Otherwise, the test data y are plotted. The plotting symbols are set by pch (the default is 1 to $\nu$), one for each group, where $\nu$ is the number of groups, and the colour by ptcol.

For 2-d plots: If display="part" then a partition induced by the discriminant analysis is also plotted. If this is not desired, set display="". Its colours are controlled by col (the default is heat.colors). Unlike plot.kde, the contour plots are automatically added to the plot. Default contours are cont=c(25,50,75). The line types are set by lty (the default is 1 to $\nu$). Also, cont and ncont control the number of level curves (only one of these needs to be set).

For 3-d plots: Default contours are cont=c(25,50). Colors are set one per group - default is heat.colors. The transparency of each contour (within each group) is alphavec. Default is seq(0.1, 0.5, length(cont)). origin is the point where the three axes meet. endpts is the vector of the maximum axis values to be plotted. Default endpts is the maxima for the plotting grid from x (automatically generated by kde).

References

Bowman, A.W. & Azzalini, A. (1997) Applied Smoothing Techniques for Data Analysis. Clarendon Press. Oxford. Simonoff, J. S., (1996) Smoothing Methods in Statistics. Springer-Verlag. New York.

See Also

kda.kde, pda.pde, kda, pda

Examples

## bivariate example

library(MASS)
data(iris)
ir <- iris[,c(1,2)]
ir.gr <- iris[,5]
xlab <- "Sepal length (mm)"
ylab <- "Sepal width (mm)"
xlim <- c(4,8)
ylim <- c(2,4.5)

H <- Hkda(ir, ir.gr, bw="plugin", pre="scale")
fhat <- kda.kde(ir, ir.gr, H)
lda.fhat <- pda.pde(ir, ir.gr, type="line")
qda.fhat <- pda.pde(ir, ir.gr, type="quad")

layout(rbind(c(1,2), c(3,4)))
plot(fhat, cont=0, xlab=xlab, ylab=ylab, xlim=xlim, ylim=ylim, 
    pch=c(1,5,10))
plot(fhat, ncont=6, xlab=xlab, ylab=ylab, xlim=xlim, ylim=ylim,
     col=c("transparent", "grey", "#8f8f8f"), drawlabels=FALSE)
plot(lda.fhat, ncont=6, xlim=xlim, ylim=ylim, xlab=xlab, ylab=ylab, 
     disp="")
plot(qda.fhat, ncont=6, xlim=xlim, ylim=ylim,  xlab=xlab, ylab=ylab, 
     lty=c(2,5,3))
layout(1)

## trivariate example
ir <- iris[,1:3]
ir.gr <- iris[,5] 
H <- Hkda(ir, ir.gr, bw="plugin", pre="scale")
fhat <- kda.kde(ir, ir.gr, H)
plot(fhat, cont=c(25,50))
   ## colour indicates species, transparency indicates density heights

qda.fhat <- pda.pde(ir, ir.gr, type="quad")
plot(qda.fhat, cont=c(25,50))