comm_detect: Community Detection Across Multiple Routines ('comm_detect')

In ideanet: Integrating Data Exchange and Analysis for Networks ('ideanet')

Community Detection Across Multiple Routines (comm_detect)

Description

The comm_detect function runs a set of several commonly-used community detection routines on a network and provides community assignments from these routines. Need to mention that only supports undirected nets and that for some routines the median community value is used.

Usage

comm_detect(g, modres = 1, slow_routines = FALSE, shiny = FALSE)

Arguments

g	An igraph object. If the igraph object contains a directed network, the function will treat the network as undirected before running community detection routines.
modres	A modularity resolution parameter used when performing community detection using the Leiden method.
slow_routines	A logical indicating whether time-intensive community detection routines should be performed on larger networks. Edge betweenness, leading eigenvector, link communities, and stochastic blockmodeling each take a very long time to identify communities in networks consisting of more than a few thousand nodes. By default, comm_detect will skip performing these routines on networks with more than 5,000 nodes and inform the user that it is doing so.
shiny	An argument indicating whether the output from the comm_detect function will be fed into the IDEANet visualization app.

Value

comm_detect returns a list contianing three data frames. comm_members indicates each node's assigned community membership from each community detection routine. comm_summaries indicates the number of communities inferred from each routine as well as the modularity score arising from community assignments. comp_scores contains a matrix indicating the similarity of community assignments between each pair of community detection routines, measured using adjusted rand scores. A fourth element in the list, plots, contains a series of network visualizations in which nodes are colored by their assigned community memberships from each routine. If shiny == FALSE, this function will display these visualizations in the user's plot window.

Examples

# Run netwrite
nw_fauxmesa <- netwrite(nodelist = fauxmesa_nodes,
                      node_id = "id",
                      i_elements = fauxmesa_edges$from,
                      j_elements = fauxmesa_edges$to,
                      directed = TRUE,
                      net_name = "faux_mesa",
                      output = "graph")

# Run comm_detect
faux_communities <- comm_detect(g = nw_fauxmesa$faux_mesa)

ideanet documentation built on April 3, 2025, 11:55 p.m.