MMD_l: Compute the MMD between two samples in linear time

In AnthonyEbert/EasyMMD: Allows users to compute the Maximum Mean Discrepancy (MMD) between two samples.

Compute the MMD between two samples in linear time

Description

Compute the MMD between two samples in linear time

Usage

MMD_l(y, x, var = 1, permute = FALSE)

Arguments

y	numeric vector, method for matrix input not yet implemented which means observations must be 1D.
x	numeric vector, method for matrix input not yet implemented which means observations must be 1D.
var	numeric kernel variance.
permute	logical Should y and x be permuted before MMD_l computation?

References

Gretton, Arthur, et al. "A kernel two-sample test." Journal of Machine Learning Research 13.Mar (2012): 723-773.

Examples

y <- rnorm(2000)
x <- rnorm(2000, 5)

MMD_full <- MMD(y, x)
MMD_linear <- MMD_l(y, x)

MMD_full
MMD_linear

AnthonyEbert/EasyMMD documentation built on March 29, 2022, 8:55 p.m.