scaleDistPlot: Annotated multidimensional scaling plots.

In clst: Classification by local similarity threshold

Description

Produces annotated representations of two-dimensional multidimensional scaling plots using cmdscale.

Usage

scaleDistPlot(dmat, groups, fill, X, O, indices = "no",
              include, display, labels,
              shuffleGlyphs = NA, key = "top",
              keyCols = 4, glyphs,
              xflip = FALSE, yflip = FALSE, ...)

Arguments

dmat	Square matrix of pairwise distances.
groups	Object coercible to a factor identifying group membership of objects corresponding to either edge of dmat.
fill	vector (logical or indices) of points to fill
X	vector of points to mark with an X
O	vector of points to mark with a circle
indices	label points with indices (all points if 'yes', or a subset indicated by a vector)
include	boolean or numeric vector of elements to include in call to cmdscale
display	boolean or numeric vector of elements to include in call to display
labels	list or data frame with parameters $i indicating indices and $text containing labels.
shuffleGlyphs	modify permutation of shapes and colors given an integer to serve as a random seed.
key	'right' (single column), 'top' (variable number of columns), or NULL for no key
keyCols	number of columns in key
glyphs	a data.frame with columns named col and pch corresponding to elements of unique(groups)
xflip	if TRUE, flip orientation of x-axis
yflip	if TRUE, flip orientation of y-axis
...	additional arguments are passed to xyplot

Value

Returns a lattice grid object.

Author(s)

Noah Hoffman

See Also

cmdscale, xyplot

Examples

data(iris)
dmat <- as.matrix(dist(iris[,1:4], method="euclidean"))
groups <- iris$Species

## visualize pairwise euclidean dstances among items in the Iris data set
fig <- scaleDistPlot(dmat, groups)
plot(fig)

## leave-one-out analysis of the classifier
loo <- lapply(seq_along(groups), function(i){
  do.call(classify, pull(dmat, groups, i))
})
matches <- lapply(loo, function(x) rev(x)[[1]]$matches)
result <- sapply(matches, paste, collapse='-')
confusion <- sapply(matches, length) > 1
no_match <- sapply(matches, length) < 1
plot(scaleDistPlot(dmat, groups, fill=confusion, O=confusion, X=no_match))

clst documentation built on Nov. 8, 2020, 5:41 p.m.