backbone (version 2.0.0)

global: Compute global threshold backbone

Description

global extracts the backbone of a weighted network using a global threshold

Usage

global(
  W,
  upper = 0,
  lower = NULL,
  keepzeros = TRUE,
  class = "original",
  narrative = FALSE
)

Arguments

W

A weighted unipartite graph, as: (1) an adjacency matrix in the form of a matrix, sparse Matrix, or dataframe; (2) an edgelist in the form of a three-column matrix, sparse Matrix, or dataframe; (3) an igraph object; (4) a network object.

upper

real, FUN, or NULL: upper threshold value or function that evaluates to an upper threshold value.

lower

real, FUN, or NULL: lower threshold value or function that evaluates to a lower threshold value.

keepzeros

boolean: TRUE if zero-weight edges in W should be excluded from (i.e. also be zero in) the backbone

class

string: the class of the returned backbone graph, one of c("original", "matrix", "sparseMatrix", "igraph", "network", "edgelist"). If "original", the backbone graph returned is of the same class as W.

narrative

boolean: TRUE if suggested text & citations should be displayed.

Value

Binary or signed backbone graph of class given in parameter class.

Details

The global function retains a edge in the backbone if its weight exceeds upper. If a lower threshold is also specified, it returns a signed backbone in which edge weights are set to 1 if above the given upper threshold, set to -1 if below the given lower threshold, and set to 0 otherwise. Prior to v2.0.0, the global function was called universal(), which is now depricated.

If W is an unweighted bipartite graph, any rows and columns that contain only zeros or only ones are removed, then the global threshold is applied to its weighted bipartite projection.

References

Neal, Z. P. (2014). The backbone of bipartite projections: Inferring relationships from co-authorship, co-sponsorship, co-attendance, and other co-behaviors. Social Networks, 39, 84-97. 10.1016/j.socnet.2014.06.001

Examples

# NOT RUN {
W <- matrix(sample(0:5, 100, replace = TRUE), 10) #Random weighted graph
diag(W) <- 0
W
global(W, narrative = TRUE)  #Keep all non-zero edges
global(W, upper = 4, lower = 2, narrative = TRUE)  #Signed with specified thresholds
global(W, upper = function(x)mean(x),  #Above-average --> positive edges
          lower = function(x)mean(x), narrative = TRUE)  #Below-average --> negative edges
# }