Permutation test based on Wasserstein distance

wasserstein_permut(
  x,
  y,
  iterations = 10^4,
  fast = nrow(x) + nrow(y) > 10^3,
  S = NULL,
  ...
)

Arguments

x	Samples from the first distribution
y	Samples from the second distribution. Only used if x is a vector.
iterations	How many iterations to do to simulate the null distribution. Default to 10^4.
fast	If true, uses the subwasserstein approximate function. Default to true if there are more than 1,000 samples total.
S	Number of samples to use in approximate mode. Must be set if `fast=TRUE`. See subwasserstein.
...	Other parameters passed to wasserstein or wasserstein1d

Value

A list containing the following components:

statistic the Wasserstein distance between x and y.
p.value the p-value of the permutation test.

Examples

 x <- matrix(c(runif(100, 0, 1),
               runif(100, -1, 1)),
             ncol = 2)
 y <- matrix(c(runif(100, 0, 3),
               runif(100, -1, 1)),
             ncol = 2)
 # Set iterations to small number for runtime
 # Increase for more accurate results
 wasserstein_permut(x, y, iterations = 10^2)
#> $statistic
#> [1] 0.9740584
#> 
#> $p.value
#> [1] 0.01
#>