BGGM (version 1.0.0)

select.estimate: Select Graphical Structure with Estimation Based Methods

Description

This allows for not only estimating the conditional dependence structure, that is non-zero edges, but also the conditional **in**dependence structure (evidence for no relation). For the latter, the region of practical equivalence must be specified

Usage

# S3 method for estimate
select(object, cred = 0.95, rope = NULL, prob = 0.95, ...)

Arguments

object

object of class estimate.default

cred

credible interval width used for the decision rule

rope

region of practical equivalence

prob

posterior probability (see notes)

...

not currently used

Value

An object of class select.estimate:

analytic = TRUE:

• partials_non_zero selected partial correlation matrix

• adjacency_non_zero adjacency matrix for the selected edges

• ci credible interval width

• analytic TRUE

• pcors_samples posterior samples

analytic = FALSE:

• partials_non_zero selected partial correlation matrix

• adjacency_non_zero adjacency matrix for the selected edges

• pcor_sd posterior standard deviation

• ci credible interval width

• rope NULL

credible interval:

• partials_non_zero selected partial correlation matrix (outside of the rope)

• adjacency_non_zero adjacency matrix for the selected edges (outside of the rope)

• partials_zero partials in the rope

• adjaceny_zero adjacency in the rope

• pcor_sd posterior standard deviation

• call match.call()

• rope specified rope

• in_rope probability in the rope

• pcors_samples posterior samples

Examples

# NOT RUN {
# Analytic = TRUE
# p = 5
Y <- BGGM::bfi[,1:5]

# analytic solution
fit_analytic <- estimate(Y, analytic = TRUE)

# select E
E <- select(fit_analytic, ci_width = 0.95)

# non-zero partial correlations
E$partials_non_zero

# adjacency matrix
E$adjacency_non_zero
# }