tang.test: Tang hypothesis testing for random graphs.

In statGraph: Statistical Methods for Graphs

Description

Given two independent finite-dimensional random dot product graphs, tang.test tests if they have generating latent positions that are drawn from the same distribution.

Usage

tang.test(G1, G2, dim, sigma = NULL, maxBoot = 200)

Arguments

G1	the first undirected graph to be compared. Must be an igraph object.
G2	the second undirected graph to be compared. Must be an igraph object.
dim	dimension of the adjacency spectral embedding.
sigma	a real value indicating the kernel bandwidth. If NULL (default) the bandwidth is calculated by the method.
maxBoot	integer indicating the number of bootstrap resamples (default is 200).

Value

A list containing:

T	the value of the test.
p.value	the p-value of the test.

References

Tang, Minh, et al. "A nonparametric two-sample hypothesis testing problem for random graphs." Bernoulli 23.3 (2017): 1599-1630.

Tang, Minh, et al. "A semiparametric two-sample hypothesis testing problem for random graphs." Journal of Computational and Graphical Statistics 26.2 (2017): 344-354.

Examples

set.seed(42)

## test under H0
lpvs <- matrix(rnorm(200), 20, 10)
lpvs <- apply(lpvs, 2, function(x) { return (abs(x)/sqrt(sum(x^2))) })
G1 <- igraph::sample_dot_product(lpvs)
G2 <- igraph::sample_dot_product(lpvs)
D1 <- tang.test(G1, G2, 5)
D1

## test under H1
lpvs2 <- matrix(pnorm(200), 20, 10)
lpvs2 <- apply(lpvs2, 2, function(x) { return (abs(x)/sqrt(sum(x^2))) })
G2 <- suppressWarnings(igraph::sample_dot_product(lpvs2))
D2 <- tang.test(G1, G2, 5)
D2

statGraph documentation built on May 19, 2021, 9:11 a.m.