plot.uncertMC: Plot method for 'uncertMC' objects

In metRology: Support for Metrological Applications

Description

Plots for uncertainty evaluations produced by uncertMC or calls to uncert with method=MC.

Usage

## S3 method for class 'uncertMC'
plot(x, which = 1:2, 
    main=paste("Monte Carlo evaluation -",deparse(substitute(x))), 
    ask = prod(par("mfcol")) < length(which) && dev.interactive(), 
    caption = list("Histogram", "Q-Q plot", "Density", 
    "Correlation x-y", "Covariance x-y"), 
    xlab = paste(deparse(substitute(x)), "$y", sep = ""), 
    ..., cex.caption = 1, cex.main = 1.25, lwd.y = 2, col.y = 2, lty.y,
    col.qqline = NULL, lty.qqline = NULL, lwd.qqline = NULL)

Arguments

x	An object of class uncertMC produced by uncertMC() or uncert() with method="MC".
which	Integer in 1:5; the particular variant(s) of plot required. A vector is permitted, in which case plots are produced in ascending order of which.
main	Main title for the plot
ask	logical; if 'TRUE', the user is _ask_ed before each plot, see 'par("ask=")'
caption	A list of captions for all the different plots.
xlab	x-axis label, currently passed only to the histogram plot.
...	Additional parameters passed to other functions. See details for which parameters are passed.
cex.caption	Expansion factor for individual plot captions; as cex in par.
cex.main	Expansion factor for main title; as cex.main in par.
lwd.y, col.y, lty.y	Line width and colour for the location line in the histogram and density plots. Setting lwd.y=0 suppresses the location line.
col.qqline, lty.qqline, lwd.qqline	Graphical parameters for the Q-Q line in the Q-Q plot.

Details

For uncert objects created with methods other than MC, the plot types are:

which=1

A histogram of the MC replicates in x$MC$y, with optional line for x$MC$y. The histogram is produced using hist.default

which=2

A Q-Q plot of the MC replicates in x$MC$y, with Q-Q line. The plot uses qqnorm.default. If datax is not present (in sQuote...), it is set to TRUE.

which=3

A density plot of the MC replicates in x$MC$y. The plot calls density.default to calculate the density and plot.density to produce the plot.

which=4

A bar plot of eqncor(x_i,y) if x$y is present. Any correlation method supported by stats::cor may be included in ‘...’ (e.g as method="pearson".

which=5

A bar plot of eqncov(x_i,y) if x$y is present. Any correlation method supported by stats::cov may be included in ‘...’ (e.g as method="pearson".

Values outside 1:5 are silently ignored.

Parameters in ‘...’ are passed to the various plot methods or calculations called. Only those parameters relevant to a given plot are passed to each calculation or plotting function, so ‘...’ can include any parameter accepted by any of the functions called.

For the x-y correlation and x-y covariance plot, values in x$cor.xy are used if available. If not, stats::cor or stats::cov is called on values in x$MC$y and x$MC$x if the latter is available (i.e. uncertMC was called with keep.x=TRUE). If neither x$cor.xy nor x$MC$x is present, or if method is unknown, the plot is skipped with a warning.

Value

plot.uncertMC invisibly returns NULL.

Author(s)

S. L. R. Ellison, [e-mail]

References

None.

See Also

uncertMC-class, hist, qqnorm, qqline, density, plot.density

Examples

  expr <- expression(a/(b-c))
  x <- list(a=1, b=3, c=2)
  u <- lapply(x, function(x) x/20)
  set.seed(403)
  u.invexpr<-uncertMC(expr, x, u, distrib=rep("norm", 3), B=999, keep.x=TRUE )
  par(mfrow=c(2,2))
  plot(u.invexpr, which=1:4, pch=20, method="k") 
                                # method="k" gives Kendall correlation 

Example output

Attaching package: ‘metRology’

The following objects are masked from ‘package:base’:

    cbind, rbind

metRology documentation built on May 2, 2019, 12:20 p.m.