Resample

Simulates observations from a multinomial distribution.

Usage

resample(object, ...)

# S4 method for numeric
resample(object, do, n, size = sum(object), ..., f = NULL)

Arguments

object

A numeric vector of count data (absolute frequencies).

...

Extra arguments passed to do.

do

A function that takes object as an argument and returns a single numeric value.

n

A non-negative integer specifying the number of bootstrap replications.

size

A non-negative integer specifying the sample size.

f

A function that takes a single numeric vector (the result of do) as argument.

Value

If f is NULL, resample() returns the n values of do. Else, returns the result of f applied to the n values of do.

See also

stats::rmultinom()

Other resampling methods: bootstrap(), jackknife()

Author

N. Frerebeau

Examples

## Sample observations from a multinomial distribution
x <- sample(1:100, 50, TRUE)
resample(x, do = median, n = 100)
#>   [1] 49.5 50.5 47.0 49.5 50.5 52.5 45.0 48.0 49.5 46.5 50.5 50.0 46.0 44.5 50.0
#>  [16] 48.5 49.0 52.0 49.5 50.5 44.5 48.5 45.0 51.0 48.5 47.5 48.5 44.0 48.5 50.0
#>  [31] 49.0 44.0 55.0 48.5 51.0 47.5 45.5 44.5 44.5 47.5 47.0 46.5 48.5 49.0 47.5
#>  [46] 51.0 45.0 49.5 44.0 46.5 47.0 49.0 44.5 44.0 46.5 46.5 50.0 51.5 43.0 43.0
#>  [61] 46.5 43.0 46.5 48.0 45.5 51.0 50.0 47.5 47.0 46.0 46.0 52.5 45.5 55.0 48.5
#>  [76] 50.0 46.0 49.5 50.5 51.0 43.0 47.5 45.5 52.5 51.0 46.0 46.0 48.0 53.5 54.5
#>  [91] 45.5 47.5 44.5 45.5 51.0 51.0 47.0 46.5 46.5 44.5

## Estimate the 25th, 50th and 95th percentiles
quant <- function(x) { quantile(x, probs = c(0.25, 0.50, 0.75)) }
resample(x, n = 100, do = median, f = quant)
#>    25%    50%    75% 
#> 46.000 47.500 49.125