fun.testboot: Compare coverage of bootstrap confidence intervals.

In kolassa-dev/NonparametricHeuristic: Heuristic Tools for Nonparametric Statistics

Compare coverage of bootstrap confidence intervals.

Description

Compare coverage of bootstrap confidence intervals.

Usage

fun.testboot(
  fcn,
  sampsize = 10,
  nsamp = NA,
  mcsamp = 10000,
  B = 9999,
  dists = c("rnorm", "rcauchy"),
  truem = NULL,
  alpha = 0.05
)

Arguments

fcn	function to be bootstrapped.
sampsize	Number of observations per sample; defaults to 10
nsamp	Deprecated version of mcsamp.
mcsamp	Number of Monte Carlo samples.
B	Bootstrap size
dists	vector of names of random number generators for sample. Defaults to normal and Cauchy. This fails for a single distribution.
truem	True parameter value; should be of length 1 or length of distribution vector.
alpha	1-confidence level.

Value

A list with two components. The first is a two-dimensional array of average interval lengths. The second is coverage levels. Dimensions of both are distribution and confidence interval technique.

Examples

fun.testboot(function(x,indices) return(c(mean(x[indices]),1)),mcsamp=100,B=999)

kolassa-dev/NonparametricHeuristic documentation built on Nov. 14, 2022, 11:25 p.m.