plotLCP: Plot local coverage probabilities to assess calibration and...
In ppernot/ErrViewLib: Library for ErrView

plotLCP

R Documentation

Plot local coverage probabilities to assess calibration and sharpness

Description

Plot local coverage probabilities to assess calibration and sharpness

Usage

plotLCP(
  E,
  U,
  ordX = NULL,
  aux = NULL,
  logX = FALSE,
  prob = c(0.95),
  nBin = NULL,
  equiPop = TRUE,
  popMin = 30,
  logBin = TRUE,
  intrv = NULL,
  slide = FALSE,
  binomCI = c("wilson", "wilsoncc", "clopper-pearson", "agresti-coull", "jeffreys"),
  plot = TRUE,
  mycols = 1:length(prob),
  xlab = "Calculated value",
  xlim = NULL,
  ylim = c(0, 1),
  title = "",
  legLoc = "bottom",
  legNcol = 3,
  label = 0,
  gPars = ErrViewLib::setgPars()
)

Arguments

`E`	(vector) Errors
`U`	(matrix) Prediction uncertainties
`ordX`	(vector) set of abscissas to order sample
`aux`	(vector) auxilliary vector to resolve ties in ordX sorting
`logX`	(logical) log-transform abscissas
`prob`	(vector) a set of coverage probabilities for the PUs
`nBin`	(integer) number of intervals for local coverage stats
`equiPop`	(logical) generate intervals with equal bin counts (default: 'equiPop = TRUE')
`popMin`	(integer) minimal bin count in an interval
`logBin`	(logical) if 'equiPop = FALSE', one can choose between equal range intervals, or equal log-range intervals (default 'logBin = TRUE')
`intrv`	(object) intervals generated by 'genIntervals' (default: 'NULL')
`slide`	(logical) use sliding window
`binomCI`	(string) name of method to estimate Binomial Proportion CI
`plot`	(logical) plot the results
`mycols`	(vector) a set of color indexes to gPars colors
`xlab`	(string) abscissa label
`xlim`	(vector) range for abscissa
`ylim`	(vector) limits of the y axis
`title`	(string) a title to display above the plot
`legLoc`	(string) location of legend (see xy.coord)
`legNcol`	(integer) number of columns for legend
`label`	(integer) index of letter for subplot tag
`gPars`	(list) graphical parameters

Value

Invisibly returns a list of LCP results. Mainly used for its plotting side effect.

Examples


  uE  = sqrt(rchisq(1000, df = 4))  # Re-scale uncertainty
  E   = rnorm(uE, mean=0, sd=uE)    # Generate errors
  plotLCP(E, cbind(0.32*uE, uE, 1.96*uE), prob=c(0.25,0.67,0.95), ordX = uE)

ppernot/ErrViewLib documentation built on June 1, 2024, 4:33 a.m.