如何使概率分布成为R中函数的参数？

Question

I have a function f in R which involves drawing many samples of the form

sample <- rnorm(k,0,1) 

where k is some integer. I would like to make as an argument of this function f the type of distribution, so I can quickly generate samples of the form

sample <- runif(k,0,1)

or other probability distributions for instance. In other words, I want to be able to write f(k,uniform) and generate the second type of sampling, and f(k,normal) for the first.

Is this possible? I'd like to avoid having to repeatedly modify the code within my functions each time I change distributions.

Answer 1

Don't know how useful this is, but:

f <- function(k,g){g(k)}

Used like f(100,runif) or f(100,rnorm)

As a variation:

f <- function(k,g,...){g(k,...)}

which would also allow things like f(100,rnorm,10,2)

Answer 2

Here's a promising, in-development implementation of what you're looking for. The distributions package is available on Github and on CRAN.

Here's an example usage:

library(distributions)

X <- Bernoulli(0.1)

random(X, 10)
#>  [1] 0 0 0 0 0 0 0 0 1 0
pdf(X, 1)
#> [1] 0.1

cdf(X, 0)
#> [1] 0.9
quantile(X, 0.5)
#> [1] 0