compare_edges function

Compare Edges Between Gaussian Graphical Models

Establish whether each of the corresponding edges are significantly different in two groups, with the de-sparsified estimator of if(!exists(".Rdpack.currefs")) .Rdpack.currefs <-new.env();Rdpack::insert_citeOnly(keys="jankova2015confidence",package="GGMncv",cached_env=.Rdpack.currefs) .

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

Arguments

• 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.

Returns

• 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

Note

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 if(!exists(".Rdpack.currefs")) .Rdpack.currefs <-new.env();Rdpack::insert_citeOnly(keys="williams_2021;textual",package="GGMncv",cached_env=.Rdpack.currefs)

demonstrated that the confidence intervals can have (severely) compromised coverage properties (whereas non-regularized methods had coverage at the nominal level).

Examples

# 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

References

if(!exists(".Rdpack.currefs")) .Rdpack.currefs <-new.env();Rdpack::insert_all_ref(.Rdpack.currefs)