calibScoresProbRef: Probabilistic reference values for calibration scores

In ppernot/ErrViewLib: Library for ErrView

Probabilistic reference values for calibration scores

Description

Probabilistic reference values for calibration scores

Usage

calibScoresProbRef(
  E,
  u,
  X = NULL,
  nBin = NULL,
  intrv = NULL,
  nMC = 1000,
  dist = c("Normal", "Uniform", "Normp4", "Laplace", "T4"),
  quant = FALSE
)

Arguments

E	(vector) errors
u	(vector) prediction uncertainties
X	(vector/matrix) input features / coordinates (default: 'NULL')
nBin	(integer) number of bins used for statistics (default: 'sqrt(length(E))'). Superseded by 'intrv$nbr' if present.
intrv	(list) intervals generated by 'genIntervals' (default: 'NULL')
nMC	(integer) sampling size (default: 1000)
dist	(string) model error distribution to generate the reference values. One of 'Normal' (default), 'Uniform', 'Normp4', 'Laplace' or 'T4'
quant	(logical) if TRUE, return 95% confidence interval limits

Value

A list containing two lists: meanRefScores is a list of mean scores with the same structure as the output of 'calibScores' and sdRefScores contains the standard deviations with the same structure

Examples

  uE = sqrt(rchisq(1000, df = 4))  # Re-scale uncertainty
  E  = rnorm(uE, mean=0, sd=uE)    # Generate errors
  calibScoresProbRef(E, uE)

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