Rnets (version 1.2.1)

signed_modularity: Robust Estimator of modularity (Gomez, et al 2009)

Description

For flexibility, x may be provided as any of the following formats: an edgelist (data.frame), a weighted adjacency matrix (square numeric matrix), an igraph object, or an rnet.* object (e.g., rnetBasic, rnetMultiStrata, etc.).

Usage

signed_modularity(x, membership, weight = NULL)

Arguments

x

A graph presented in of the forms discussed below.

membership

Defines vertex membership to determine if vertices are similar. May be provided as a string that matches an attribute of x or a vector of length equal to the number of vertices in the graph.

weight

Edge weights. Like 'membership', this argument can be defined as a string matching an edge attribute of 'x' or a vector of length equal to the number of edges, but may also be left as NULL which will return an unweighted modularity estimate.

Value

a numeric value estimating the weighted, signed modularity of x, or a numeric vector containing respective modularity estimates if x contained multiple network.

Details

Newman's method of estimating graphical modularity based on vertex can accomodate edge weights, but cannot incorporate signed edges, e.g. edges with both positive and negative. Gomez, et al, proposed a similar estimator of modularity estimated in two parts corresponding to positive (Q+) and negative (Q-) edges, and the latter is subtracted from the former. The 'signed_modularity' function implements this method of modularity estimation, and returns a scalar.

Examples

# NOT RUN {
#Signed modularity in a random graph with 10 vertices

x <- sample_gnp(5, 0.4)  #Creates a random graph with 10 vertices and density ~ 40%
x <- set_edge_attr(x, 'weight', value = runif(gsize(x), -0.5, 0.5))  
     #Randomly assign edge weights to edge attribute 'weight', both positive and negative
x <- set_vertex_attr(x, name = 'group', value = sample(c('red', 'blue'), size = 5, replace = TRUE))

signed_modularity(x, membership = 'group', weight = 'weight')
signed_modularity(x, membership = 'group')

# }