compare_edges: Compare Edges Between Gaussian Graphical Models
In GGMncv: Gaussian Graphical Models with Nonconvex Regularization

Description Usage Arguments Value Note References Examples

Establish whether each of the corresponding edges are significantly different in two groups, with the de-sparsified estimator of \insertCitejankova2015confidenceGGMncv.

1	compare_edges(object_1, object_2, method = "fdr", alpha = 0.05, ...)

`object_1`	object of class `ggmncv` .
`object_2`	An object of class `ggmncv`.
`method`	Character string. A correction method for multiple comparisons (defaults to `fdr`), which can be abbreviated. See p.adjust.
`alpha`	Numeric. Significance level (defaults to `0.05`).
`...`	Currently ignored.

P_diff De-sparsified partial correlation differences
adj Adjacency matrix based on the p-values.
pval_uncorrected Uncorrected p-values
pval_corrected Corrected p-values
method The approach used for multiple comparisons
alpha Significance level

For low-dimensional settings, i.e., when the number of observations far exceeds the number of nodes, this function likely has limited utility and a non regularized approach should be used for comparing edges (see for example GGMnonreg).

Further, whether the de-sparsified estimator provides nominal error rates remains to be seen, at least across a range of conditions. For example, the simulation results in \insertCitewilliams_2021;textualGGMncv demonstrated that the confidence intervals can have (severely) compromised coverage properties (whereas non-regularized methods had coverage at the nominal level).

\insertAllCited

# data
# note: all edges equal
Y1 <- MASS::mvrnorm(250, rep(0, 10), Sigma = diag(10))
Y2 <- MASS::mvrnorm(250, rep(0, 10), Sigma = diag(10))

# fit models
# note: atan penalty by default

# group 1
fit1 <- ggmncv(cor(Y1), n = nrow(Y1),
               progress = FALSE)

# group 2
fit2 <- ggmncv(cor(Y2), n = nrow(Y2),
               progress = FALSE)

# compare
compare_ggms <- compare_edges(fit1, fit2)

compare_ggms