zipfR (version 0.6-70)

plot.vgc: Plot Vocabulary Growth Curves (zipfR)

Description

Plot a vocabulary growth curve (i.e., \(V(N)\) or \(V_m(N)\) against \(N\)), or a comparison of several vocabulary growth curves.

Usage

# S3 method for vgc
plot(x, y, ...,
     m=NULL, add.m=NULL, N0=NULL,
     conf.level=.95, conf.style=c("ticks", "lines"),
     log=c("", "x", "y", "xy"),
     bw=zipfR.par("bw"), 
     xlim=NULL, ylim=NULL,
     xlab="N", ylab="V(N)", legend=NULL,
     main="Vocabulary Growth",
     lty=NULL, lwd=NULL, col=NULL)

Arguments

x, y, ...

one or more objects of class vgc, representing observed or expected vocabulary growth curves to be plotted. Alternatively, all VGCs can be passed as a list in the x argument if the method is called explicitly (see ‘Examples’.

m

a single integer \(m\) in the range \(1 \ldots 9\). If specified, graphs will be plotted for \(V_m(N)\) instead of \(V(N)\) (the default). Note that all vgc objects to be plotted must contain the necessary data in this case.

add.m

a vector of integers in the range \(1 \ldots 9\). If specified, graphs for \(V_m(N)\) will be added as thin lines to the default \(V(N)\) curve, for all specified frequency classes \(m\). This option cannot be combined with the m option above. See "Details" below.

N0

if specified, draw a dashed vertical line at \(N=N_0\), indicating the sample size where a LNRE model has been estimated (this is never done automatically)

log

a character string specifying the axis or axes for which logarithmic scale is to be used ("x", "y", or "xy"), similar to the log argument of plot.default. By default, both axes use linear scale (also see "Details" below).

conf.level

confidence level for confidence intervals around expected vocabulary growth curves (see "Details" below). The default value of \(.95\) produces 95%-confidence intervals. Set to NA in order to suppress confidence interval markers.

conf.style

if "ticks", confidence intervals are indicated by vertical lines at each data point in the vgc object (default). If "lines", confidence intervals are indicated by thin curves above and below the VGC (which may be difficult to see when plotting multiple VGCs). Notice that confidence intervals might be so narrow as to be invisible in plots (one way to visualize them in such case might be to set an extremely conservative confidence level, such as \(.9999\)).

bw

if TRUE, draw plot in B/W style (default is the global zipfR.par setting)

xlim, ylim

visible range on x- and y-axis. The default values are automatically determined to fit the selected data in the plot.

xlab, ylab

labels for the x-axis and y-axis. The default values nicely typeset mathematical expressions. The y-axis label also distinguishes between observed and expected vocabulary growth curves, as well as between \(V(N)\) and \(V_m(N)\).

main

a character string or expression specifying a main title for the plot

legend

optional vector of character strings or expressions, specifying labels for a legend box, which will be drawn in the lower right-hand corner of the screen. If legend is given, its length must correspond to the number of VGCs in the plot.

lty, lwd, col

style vectors that can be used to override the global styles defined by zipfR.par. If these vectors are specified, they must contain at least as many elements as there are VGCs in the plot: the values are not automatically recycled.

Details

By default, standard vocabulary growth curves are plotted for all specified vgc objects, i.e. graphs of \(V(N)\) against \(N\). If m is specified, growth curves for hapax legomena or other frequency classes are shown instead, i.e. graphs of \(V_m(N)\) against \(N\). In this case, all vgc objects must contain the necessary data for \(V_m(N)\).

Alternatively, the option add.m can be used to display growth curves for one or more spectrum elements in addition to the standard VGCs. These growth curves are plotted as thinner lines, otherwise matching the styles of the main curves. Since such plots can become fairly confusing and there is no finer control over the styles of the additional curves, it is generally not recommended to make use of the add.m option.

Confidence intervals are indicated for expected vocabulary growth curves with variance data, either by short vertical lines (conf.style="ticks", the default) or by thin curves above and below the main growth curve (conf.style="lines"). The size of the confidence intervals is controlled by the conf.level parameter (default: 95%). Set conf.level=NA in order to suppress the confidence interval indicators.

In y-logarithmic plots, data points with \(V(N) = 0\) or \(V_m(N) = 0\) are drawn outside the plot region (below the bottom margin) rather than skipped.

Line and point styles can be defined globally with zipfR.par. They can be overridden locally with the optional parameters lty, lwd and col, but this should only be used when absolutely necessary. In most cases, it is more advisable to change the global settings temporarily for a sequence of plots.

The bw parameter is used to switch between B/W and color modes. It can also be set globally with zipfR.par.

See Also

vgc, lnre, lnre.vgc, plot.tfl, plot.spc, zipfR.par, zipfR.plotutils

Examples

# NOT RUN {
## load Our Mutual Friend spectrum and empirical vgc
data(DickensOurMutualFriend.emp.vgc)
data(DickensOurMutualFriend.spc)

## plot empirical V and V1 growth
plot(DickensOurMutualFriend.emp.vgc,add.m=1)

## use log scale for y-axis
plot(DickensOurMutualFriend.emp.vgc,add.m=1,log="y")

## binomially interpolated vgc at same points as
## empirical vgc
omf.bin.vgc <- vgc.interp(DickensOurMutualFriend.spc,N(DickensOurMutualFriend.emp.vgc))

## compare empirical and interpolated vgc, also with
## thinner lines, and in black and white
plot(DickensOurMutualFriend.emp.vgc,omf.bin.vgc,legend=c("observed","interpolated"))
plot(DickensOurMutualFriend.emp.vgc,omf.bin.vgc,legend=c("observed","interpolated"),lwd=c(1,1)) 
plot(DickensOurMutualFriend.emp.vgc,omf.bin.vgc,legend=c("observed","interpolated"),bw=TRUE)


## load Great Expectations spectrum and use it to
## compute ZM model
data(DickensGreatExpectations.spc)
ge.zm <- lnre("zm",DickensGreatExpectations.spc)

## expected V of Great Expectations at sample
## sizes of OMF's interpolated vgc
ge.zm.vgc <- lnre.vgc(ge.zm,N(omf.bin.vgc))

## compare interpolated OMF Vs and inter/extra-polated
## GE Vs, with a vertical line at sample size
## used to compute GE model
plot(omf.bin.vgc,ge.zm.vgc,N0=N(ge.zm),legend=c("OMF","GE"))


## load Italian ultra- prefix data and compute zm model
data(ItaUltra.spc)
ultra.zm <- lnre("zm",ItaUltra.spc)

## compute vgc up to about twice the sample size
## with variance of V
ultra.zm.vgc <- lnre.vgc(ultra.zm,(1:100)*70, variances=TRUE)

## plot with confidence intervals derived from variance in
## vgc (with larger datasets, ci will typically be almost
## invisible)
plot(ultra.zm.vgc)

## use more conservative confidence level, and plot 
## the intervals as lines
plot(ultra.zm.vgc,conf.level=.99,conf.style="lines")

## suppress ci plotting, and insert different title and labels
plot(ultra.zm.vgc,conf.level=NA,main="ultra-",xlab="sample sizes",ylab="types")

## load Brown adjective spectrum
## (about 80k tokens) 
data(BrownAdj.spc)

## binomially interpolated curve of V and V_1 to V_5
BrownAdj.bin.vgc <- vgc.interp(BrownAdj.spc,(1:100)*800,m.max=5)

## plot with V and 5 spectrum elements
plot(BrownAdj.bin.vgc,add.m=c(1:5))

## can pass list of VGCs in first argument with explicit call
plot.vgc(lapply(EvertLuedeling2001, vec2vgc), 
         xlim=c(0, 30000), ylim=c(0, 1200),
         legend=names(EvertLuedeling2001))
# }