squared_wass_decomp: Compute the squared 2-Wasserstein distance based on a...

In goncalves-lab/waddR: Statistical tests for detecting differential distributions based on the 2-Wasserstein distance

Compute the squared 2-Wasserstein distance based on a decomposition

Description

Computes the squared 2-Wasserstein distance between two vectors based on a decomposition into location, size and shape terms. For a detailed description of the (empirical) calculation of the invoved quantities, see Schefzik et al. (2020).

Usage

squared_wass_decomp(x, y)

Arguments

x	sample (vector) representing the distribution of condition A
y	sample (vector) representing the distribution of condition B

Value

A list of 4:

• distance: the sum location+size+shape

• location: location part in the decoposition of the 2-Wasserstein distance

• size: size part in the decoposition of the 2-Wasserstein distance

• shape: shape part in the decoposition of the 2-Wasserstein distance

References

Schefzik, R., Flesch, J., and Goncalves, A. (2020). waddR: Using the 2-Wasserstein distance to identify differences between distributions in two-sample testing, with application to single-cell RNA-sequencing data.

See Also

See the functions wasserstein_metric and squared_wass_approx for alternative implementations of the 2-Wasserstein distance

Examples

set.seed(24)
x<-rnorm(100)
y1<-rnorm(150)
y2<-rexp(150,3)
y3<-rpois(150,2)

squared_wass_decomp(x,y1)
squared_wass_decomp(x,y2)
squared_wass_decomp(x,y3)

goncalves-lab/waddR documentation built on June 29, 2023, 12:18 a.m.