eo.setup {UCS} | R Documentation |
eo.setup
initialises a new (e,o) plot window, which can then be
drawn into with calls to eo.iso
, eo.points
and similar
functions. The plot has to be finalised with eo.close
before a
new plot can be generated.
A detailed explanation of (e,o) plots and their interpretation can be found in Section 3.3 of Evert (2004).
eo.setup(xlim=eo.par("xlim"), ylim=eo.par("ylim"), ds=NULL, bw=eo.par("bw"), file=NULL, aspect=eo.par("aspect"), log.marks=eo.par("log.marks"), cex=eo.par("cex"), ...) eo.close()
xlim, ylim |
integer vectors of length 2, specifying ranges for
the e-axis (xlim ) and o-axis (ylim ) in orders of
magnitude (i.e., base 10 logarithms: -2 corresponds to .01, 0
corresponds to 1, and 3 corresponds to 1000). If xlim and
ylim are not given and no default values have been set with
eo.par , the ds parameter has to be specified. Note
that (e,o) plots are always drawn in logarithmic scale. |
ds |
A data set from which suitable ranges for the e-axis and
o-axis are computed. The automatically determined values are
overridden by explicit xlim and ylim parameters. |
bw |
If TRUE , the (e,o) plot is drawn in B/W mode,
otherwise in colour mode. The default value is set with
eo.par . |
file |
a character string giving the name of a PostScript
file. If specified, the (e,o) plot is saved to file in EPS
format rather than displayed on screen. Note that this file will
only be written after eo.close has been called. |
aspect |
If TRUE , an aspect ratio of 1:1 is enforced by
extending the axis ranges as necessary (assuming that the plotting
region is square). The default value is set with eo.par . |
log.marks |
If TRUE , tick marks on the axes are labelled
in logarithmic units, i.e. orders of magnitude. Otherwise, absolute
numbers are used. The default value is set with eo.par .
(Recall that (e,o) plots are always drawn in logarithmic scale.) |
cex |
overall character expansion factor (for tick marks, axis
labels and legends). The default value is set with eo.par . |
... |
Any additional parameters are passed through to the
plot function used to set up the plot region and axes. |
An (e,o) plot is typically created in four stages:
eo.setup
, defining suitable ranges
for the e-axis. These ranges and some other state information
(e.g. whether the plot is drawn in colour or B/W mode) are recorded
in the global variable .eo.STATE
.
eo.points
and
iso-lines for GAMs with eo.iso
. Differences between two
acceptance regions can be highlighted with eo.iso.diff
. The
eo.mark
function can be used to mark individual points with
circles.
eo.legend.points
, eo.legend.iso
and
eo.legend.diff
.
eo.close
. When a file
argument has been specified in the eo.setup
call, the plot
will be saved to a PostScript file at this stage.
Default values for xlim
, ylim
, bw
, aspect
,
log.marks
and cex
can be set with the eo.par
function. See the eo.par
help page for "factory settings" of
these parameters, as well as default line and point styles in colour
and B/W mode.
Note that (e,o) plots are always drawn in logarithmic scale and tick
marks are shown for orders of magnitude (full powers of ten). The
log.marks
parameter only determines whether the labels on these
tick marks show linear (.1, 1, 10, 100, ...) or logarithmic (-1, 0,
1, 2, ...) values.
Evert, Stefan (2004). The Statistics of Word Cooccurrences: Word Pairs and Collocations. PhD Thesis, IMS, University of Stuttgart.
eo.par
, eo.points
, eo.iso
,
eo.iso.diff
, eo.mark
,
eo.legend
ucs.library("eo") ## load data set file, add jitter, and reduce to random 10% subset ds <- add.jitter(read.ds.gz("dickens.ds.gz")) select <- rbinom(nrow(ds), 1, .1) == 1 ds <- ds[select,] ## 1) set up new (e,o) plot with suitable axis ranges eo.setup(ds=ds) # note that y axis is extend to enforce 1:1 aspect ## 2) add data set as point cloud and three iso-lines eo.points(ds, style=5, jitter=TRUE) eo.iso("Poisson.pv", 3, style=1) # p-value = 1e-3 eo.iso("z.score", 3.09, style=2) # corresponding one-sided z-score eo.iso("t.score", 3.09, style=3) # same as t-score with df=Inf ## 3) add legend boxes in top right (2) and bottom right (3) corner eo.legend.points(2, "pair type", 5) eo.legend.iso(3, c("Poisson", "z-score", "t-score"), 1:3) ## 4) finalise the (e,o) plot eo.close()