Jackknife Estimation — jackknife • arkhe

Jackknife Estimation

Usage

jackknife(object, ...)

# S4 method for class 'numeric'
jackknife(object, do, ..., f = NULL)

Arguments

object: A numeric vector.
...: Extra arguments to be passed to do.
do: A function that takes object as an argument and returns a single numeric value.
f: A function that takes a single numeric vector (the leave-one-out values of do) as argument.

Value

If f is NULL (the default), jackknife() returns a named numeric vector with the following elements:

original: The observed value of do applied to object.
mean: The jackknife estimate of mean of do.
bias: The jackknife estimate of bias of do.
error: he jackknife estimate of standard error of do.

If f is a function, jackknife() returns the result of f applied to the leave-one-out values of do.

See also

Other resampling methods: bootstrap()

Author

N. Frerebeau

Examples

x <- rnorm(20)

## Bootstrap
bootstrap(x, do = mean, n = 100)
#>   original       mean       bias      error 
#> 0.16598229 0.16920507 0.00322278 0.21290410 

## Estimate the 25th and 95th percentiles
quant <- function(x) { quantile(x, probs = c(0.25, 0.75)) }
bootstrap(x, n = 100, do = mean, f = quant)
#>        25%        75% 
#> 0.04353926 0.31796223 

## Get the n bootstrap values
bootstrap(x, n = 100, do = mean, f = function(x) { x })
#>   [1] -0.134917467  0.253760075 -0.064675771 -0.140809858  0.455947373
#>   [6]  0.213983815  0.196210451  0.099908964 -0.296733887 -0.119850799
#>  [11]  0.118366473 -0.284519619 -0.146630253  0.118460533  0.076898099
#>  [16]  0.539367493  0.218444817  0.213597934 -0.042169090  0.360466541
#>  [21] -0.124255012 -0.132997955  0.151614276  0.228940981  0.003373803
#>  [26]  0.092416532  0.206588480  0.306116562  0.120925361  0.120793195
#>  [31]  0.184290522  0.053559998  0.476776902 -0.246857199  0.118787874
#>  [36]  0.018281825  0.009212510  0.230146604  0.208306334  0.126052441
#>  [41]  0.365973626  0.298099535 -0.010088778  0.036630307 -0.022314473
#>  [46]  0.419389900  0.265151228 -0.088192478  0.095860729  0.160796198
#>  [51] -0.012741747 -0.068974846 -0.104448217  0.258792012  0.140757269
#>  [56]  0.393412777  0.084320022  0.182078560  0.135929209  0.557696283
#>  [61]  0.224675605 -0.199221549  0.023791573  0.337681331  0.351483128
#>  [66]  0.212835512  0.136610134  0.298267439  0.346560244  0.180953742
#>  [71]  0.354643769 -0.061265730 -0.086797454  0.083831877  0.098110049
#>  [76] -0.064267806  0.290736918  0.360256437 -0.054748577 -0.102948876
#>  [81]  0.014794758  0.143079499 -0.010388815  0.233308891  0.022045599
#>  [86]  0.017547565  0.117126863  0.061413693  0.404698885 -0.153191689
#>  [91]  0.242601159  0.280123509 -0.067535924 -0.060736928  0.012698300
#>  [96]  0.184779716  0.016434007 -0.313326247 -0.015372042  0.236452785

## Jackknife
jackknife(x, do = mean) # Sample mean
#>  original      mean      bias     error 
#> 0.1659823 0.1659823 0.0000000 0.2126934 

## Get the leave-one-out values instead of summary
jackknife(x, do = mean, f = function(x) { x })
#>  [1] 0.19818477 0.20310250 0.20462032 0.12276561 0.28212782 0.18031634
#>  [7] 0.08386451 0.15273009 0.19030649 0.12084963 0.18365068 0.22070922
#> [13] 0.16211979 0.17601232 0.12953007 0.10611419 0.19517462 0.17605627
#> [19] 0.16183668 0.06957391