Skip to contents

Edge Betweenness Network

Source: R/centrality_measures.R
net_edge_betweenness.Rd

Builds a network in which each edge's weight is replaced by its betweenness: the number of shortest paths between all node pairs that traverse that edge (fractional when shortest paths tie). This is the Nestimate counterpart of tna::betweenness_network() and produces identical values for transition networks; the name differs to avoid a clash with tna::betweenness_network() and igraph::edge_betweenness().

Usage

net_edge_betweenness(x, invert = TRUE, ...)

# S3 method for class 'netobject'
net_edge_betweenness(x, invert = TRUE, ...)

# S3 method for class 'netobject_group'
net_edge_betweenness(x, invert = TRUE, ...)

# Default S3 method
net_edge_betweenness(x, invert = TRUE, ...)

# S3 method for class 'net_edge_betweenness'
plot(x, style = c("bar", "forest", "delta"), top_n = NULL, labels = TRUE, ...)

Arguments

x

A netobject or netobject_group. For the plot() method: an object of class net_edge_betweenness.

invert

Logical. Invert weights to distances by 1/w before computing shortest paths? Default TRUE (correct for probability and frequency networks).

...

Additional arguments (ignored). In plot.net_edge_betweenness(): Additional arguments (ignored).

style

Plot style. "bar" (default) draws one horizontal bar per edge; "forest" draws a forest/lollipop chart (a stem from zero to a point) with a dashed reference line at the mean betweenness; "delta" draws each edge's deviation from the mean edge betweenness as a diverging bar (above the mean in blue, below in red).

top_n

Integer or NULL. Keep only the top_n highest edges. Default NULL (all edges with non-zero betweenness).

labels

Logical. Print the betweenness value beside each edge. Default TRUE.

Value

For a netobject: a new network of class c("net_edge_betweenness", "netobject", "cograph_network") whose $weights are the edge-betweenness scores, with method = "edge_betweenness". Call extract_edges() on it for a tidy per-edge table, or plot() to render it. The object preserves source-network metadata so permutation can test edge-betweenness differences by permuting the source networks. For a netobject_group: a netobject_group of such networks, one per group.

In plot.net_edge_betweenness(): A ggplot object.

Details

For a probability/transition network the edge weights are transition probabilities, so they are inverted to distances (invert = TRUE) before path-finding: the geodesic between two states is then the most probable route rather than the one with the fewest hops. Pass invert = FALSE when the weights already represent distances.

Directedness is taken from the network itself. A directed network yields an asymmetric betweenness matrix; an undirected (symmetric) network yields a symmetric one.

Methods

  • plot.net_edge_betweenness(): Draws the edges of a net_edge_betweenness network ranked by their betweenness, as a horizontal bar chart. This is the tidy, cograph-free companion to the node-link diagram: render the diagram with cograph::splot(eb) and the ranking with plot(eb).

Examples

seqs <- data.frame(
  V1 = c("A","B","A","C"), V2 = c("B","C","B","A"),
  V3 = c("C","A","C","B"))
net <- build_network(seqs, method = "relative")
eb  <- net_edge_betweenness(net)
extract_edges(eb)
#>   from to weight
#> 1    C  A      3
#> 2    A  B      3
#> 3    B  C      3
#> 4    B  A      0
#> 5    C  B      0
#> 6    A  C      0