plot_dist_matrix: Plotting Distance Matrices (wordspace)

In wordspace: Distributional Semantic Models in R

Plotting Distance Matrices (wordspace)

Description

Visualization of a DSM distance matrix as a neighbourhood graph based on multidimensional scaling (MDS).

Usage

  ## S3 method for class 'dist.matrix'
plot(x, y, labels=rownames(x), show.labels=TRUE, label.pos=3,
     selected=attr(x, "selected"), show.selected=TRUE,
     col="black", cex=1, pch=20, pt.cex=1.2, selected.cex=1.2, selected.col="red",
     show.edges=TRUE, edges.lwd=6, edges.col="#AABBFF", edges.threshold=quantile(x, 2/3),
     method=c("isomds", "sammon"), aspect=1, expand=.05, ...)

Arguments

x	a symmetric distance matrix of class dist.matrix. NB: similarity values and asymmetric distance measures are not supported.
y	unused, must not be specified
labels	a character vector of labels for the DSM vectors (defaults to rownames of x)
show.labels	if TRUE (default), labels are displayed if available
label.pos	position of labels (default: above points). Possible values are 1 (below), 2 (left), 3 (above) and 4 (right).
selected	logical vector of selected points that will be highlighted (defaults to optional selected attribute of distance matrix)
show.selected	if TRUE (default), points marked by selected are highlighted in the plot
col	colour of points and labels
cex	numeric character expansion factor for points and labels
pch	plot symbol for points
pt.cex	character expansion factor for points relative to labels
selected.cex	additional character expansion factor for selected points and labels
selected.col	colour of selected points and labels (if show.selected=TRUE)
show.edges	if TRUE (default), edges are drawn between points. The line width of each edge is proportional to the distance between the corresponding points.
edges.lwd	maximal line width of edges (for d = 0)
edges.col	colour of edges, usually a light or translucent shade
edges.threshold	maximal distance up to which edges are drawn. The default is to display two thirds of all edges.
method	whether to perform non-metric (isomds) or metric (sammon) multidimensional scaling
aspect	aspect ratio of plot window (e.g. aspect=16/10 for a window that is 8 inches wide and 5 inches high). Setting a correct aspect ratio ensures that the distances between points in the MDS map are correctly represented in the plot.
expand	fraction by which plotting region is extended on each side. Adjust this parameter so that points and labels are completely visible.
...	all other arguments are passed to the initial plot function, which sets up the display but does not draw any graphical elements

Details

For multidimensional scaling (MDS), the functions isoMDS and sammon from the MASS package are used.

Value

Invisibly returns a two-column matrix with MDS coordinates of all displayed points and labels as rownames (if available).

Author(s)

Stephanie Evert (https://purl.org/stephanie.evert)

See Also

nearest.neighbours, which produces distance matrices suitable for plotting if the option dist.matrix=TRUE is specified

Examples

## Not run: 
plot(nearest.neighbours(DSM_Vectors, "walk_V", n=20, dist.matrix=TRUE))

## End(Not run)

wordspace documentation built on Sept. 9, 2022, 3:04 p.m.