Sampling from a Gaussian kernel density estimate in R.

kde_sampling_gaussian_kernel.R

## Make a function rkde that samples from a kernel density

#' Sample from a kernel density.
#' 
#' @param n Number of observations to sample.
#' @param x The data from which the estimate is to be computed.
#' @param bw Desired bandwidth.
#' @return A numeric vector with n sampled data points from the kernel
#' density estimator.

rkde = function(n, x, bw) rnorm(n, mean = sample(x, n, replace = TRUE), sd = bw)

## Take a data set and find a suitable bandwidth with density.
data = airquality$Wind
bw = density(data)$bw 

## Check that it works.
set.seed(313)
n = 1000000
samples = rkde(n, data, bw)
hist(samples, breaks = 1000, freq = FALSE)
lines(density(data), lwd = 2, col = "red")