EGAnet (version 2.4.0)

plot_clusters: Plot Clustered Individual Networks

Description

Visualize clusters of individual networks identified by infoCluster (or another clustering method) applied to a dynEGA object

Provides visualization modes:

• "population": Cluster-level "population" networks, obtained by stacking derivatives across individuals and estimating a new EGA

• "average": Cluster-average networks, obtained by averaging adjacency matrices across individuals in the cluster

Usage

plot_clusters(
  dynEGA.object,
  clustering,
  include,
  type = c("population", "average"),
  node.size = 3,
  label.size = 1.2,
  ...
)

Value

Plots created using compare.EGA.plots:

"population"

A plot combining cluster-level population network plots

"average"

A plot combining cluster-average network plots

Arguments

dynEGA.object

A dynEGA or a dynEGA.ind.pop object

clustering

Vector (length of individuals). A vector of cluster membership for each individual in the dynEGA.object. A common and easy option is to use an infoCluster object containing cluster assignments for individuals (see infoCluster). Accepts any vector of memberships

include

Numeric vector. Specifies which cluster memberships should be explicitly included in the plot. By default, all clusters are shown. Use this argument to restrict the visualization to a subset of clusters (e.g., `include = c(1, 3, 5)` will only display clusters 1, 3, and 5)

type

Character (length = 1). Type of visualization to produce:

• "population" (default) --- Cluster-level population networks by stacking derivatives across individuals in the cluster and re-estimating with EGA

• "average" --- Averaged networks with a consensus clustering matrix (pairwise) membership similarity that is supplied to community.detection

In either type, the argument setting applied in your `dynEGA.object` will be carried forward into the network estimation (`"population"`) and community detection (both)

node.size

Numeric (length = 1 or number of variables). Node size in network visualizations. Defaults to 3

label.size

Numeric (length = 1 or number of nodes. Label size in network visualizations. Defaults to 1.2

...

Additional arguments passed to EGA and compare.EGA.plots

Author

Hudson Golino <hfg9s at virginia.edu>

Details

This function provides flexible visualization of individual clusters obtained from infoCluster (or other clustering input; see examples):

• Cluster-level population networks treat each cluster as a "mini-population," stacking derivatives across individuals to re-estimate an EGA

• Average networks show the mean connectivity structure for each cluster, with consensus clustering memberships

Function automatically extracts the appropriate derivatives ("population") and networks ("average") from the dynEGA.object results.

See Also

dynEGA, infoCluster, EGA, compare.EGA.plots

Examples

# Load data
data <- sim.dynEGA

if (FALSE) {
# Run dynEGA with optimization
optimized_all <- dynEGA(
  data = data, level = c("individual", "population"),
  n.embed = 3:25, n.embed.optimize = TRUE
)

# Cluster individuals
clust <- infoCluster(dynEGA.object = optimized_all)

# Cluster-level population networks
plot_clusters(
  dynEGA.object = optimized_all,
  clustering = clust,
  type = "population"
)

# Average networks per cluster
plot_clusters(
  dynEGA.object = optimized_all,
  clustering = clust,
  type = "average"
)

# Cluster-level population networks, including only Cluster 2:
plot_clusters(
  dynEGA.object = optimized_all,
  clustering = clust, include = 2,
  type = "population"
)

# Using alternative clusters
plot_clusters(
  dynEGA.object = optimized_all,
  clustering = rep(1:2, each = 50), # vector of memberships
  type = "population"
)

# Run with non-optimized dynEGA
standard_all <- dynEGA(
  data = data,
  level = c("individual", "population")
)

# Obtain clusters
clust_standard <- infoCluster(dynEGA.object = standard_all)

# Plot clusters with population
plot_clusters(
  dynEGA.object = standard_all,
  clustering = clust_standard,
  type = "population"
)}