Tables of a hypergraph community ensemble

Reads, prints and plots the result of hg_communities() on a hypergraph.

Usage

# S3 method for class 'hg_communities'
print(x, n = 10L, ...)

# S3 method for class 'hg_communities'
hg_get(
  x,
  what = c("medoid", "partitions", "runs", "sizes", "ami", "ari", "nmi", "weights"),
  ...,
  converged = NULL,
  sort_by = NULL,
  top = NULL
)

# S3 method for class 'hg_communities'
plot(x, ...)

Arguments

x

An hg_communities object.

n

Number of rows of the default table to print. Default 10.

...

For plot(), additional arguments passed to cograph::splot(); otherwise unused.

what

Table to return: "medoid" (default: one row per node with its community in the AMI-medoid run), "partitions", "runs", "sizes", "ami", "ari", "nmi", or (IRMM fits only) "weights". The three similarity tables have one row per distinct pair of runs (run_a, run_b, and the similarity), without the diagonal.

converged

For what = "runs" of an IRMM fit: TRUE keeps the runs whose weights settled, FALSE the runs that reached max_iter. NULL (default) keeps every run.

sort_by

For what = "runs": a numeric column of the runs table ("modularity", "n_communities" or "iterations") to order the runs by, largest first, ties broken by run number. NULL (default) keeps run order.

top

NULL (default, every row) or the number of first rows of any table to return, applied after converged and sort_by.

Value

hg_get(): a base data.frame. Raises hypergraphs_bad_input for converged or sort_by with a table other than "runs", or converged with an Infomap fit. print(): x, invisibly. plot(): the cograph plot of the projection, coloured by the medoid communities.

Examples

dat <- data.frame(
  member = c("a", "b", "c", "a", "b", "c", "x", "y", "z", "x", "y", "z"),
  edge = rep(paste0("e", 1:4), each = 3)
)
h <- group_hypergraph(dat, "member", "edge")
if (requireNamespace("igraph", quietly = TRUE)) {
  fit <- hg_communities(h, n_runs = 2, trials = 2, seeds = 1:2)
  hg_get(fit, what = "runs")
  hg_get(fit, what = "runs", sort_by = "n_communities", top = 1)
}
#>   run seed n_communities codelength
#> 1   1    1             2   2.052397