Per-iteration edge counts for convergence checking

Computes the number of edges in each group at each saved iteration. Useful for trace plots of model complexity to assess MCMC convergence.

Usage

edge_counts(fit, chain = 1L)

Arguments

fit

A multiggm_fit or multiggm_fit_list object.

chain

Integer; which chain to use if fit is a multiggm_fit_list. Default 1.

Value

An integer matrix with nsave rows and K columns. Entry [s, k] is the number of edges (upper triangle only) in group k at saved iteration s. Column names are "Group_1", "Group_2", etc.

See also

plot_trace(), pip_edges()

Examples

sim <- simulate_multiggm(K = 2, p = 8, n = 80, seed = 1)
fit <- multiggm_mcmc(data_list = sim$data_list, burnin = 200, nsave = 100)
ec <- edge_counts(fit)
plot(ec[,1], type = "l", ylab = "Edges", main = "Group 1 edge count trace")