Fitting gHypEG models

ghype is used to fit gHypEG models when the propensity matrix is known. It can be used to estimate a null model (soft configuration model), or the benchmark 'full-model', where the propensity matrix is fitted such that the expected graph from the fitted model is the one passed to the function.

ghype(
  graph,
  directed,
  selfloops,
  xi = NULL,
  omega = NULL,
  unbiased = FALSE,
  regular = FALSE,
  ...
)

# S3 method for matrix
ghype(
  graph,
  directed,
  selfloops,
  xi = NULL,
  omega = NULL,
  unbiased = FALSE,
  regular = FALSE,
  ...
)

# S3 method for default
ghype(
  graph,
  directed,
  selfloops,
  xi = NULL,
  omega = NULL,
  unbiased = FALSE,
  regular = FALSE,
  ...
)

# S3 method for igraph
ghype(
  graph,
  directed,
  selfloops,
  xi = NULL,
  omega = NULL,
  unbiased = FALSE,
  regular = FALSE,
  ...
)

# S3 method for ghype
print(x, suppressCall = FALSE, ...)

Arguments

graph

either an adjacency matrix or an igraph graph.

directed

a boolean argument specifying whether graph is directed or not.

selfloops

a boolean argument specifying whether the model should incorporate selfloops.

xi

an optional matrix defining the combinatorial matrix of the model.

omega

an optional matrix defining the propensity matrix of the model.

unbiased

a boolean argument specifying whether to model the hypergeometric ensemble (no propensity), defaults to FALSE.

regular

a boolean argument specifying whether to model the 'gnp' ensemble (no xi), defaults to FALSE.

...

further arguments passed to or from other methods.

x

ghype model

suppressCall

boolean, suppress print of the call

Value

ghype return an object of class "ghype".

Methods (by class)

• ghype(matrix): Fitting ghype models from an adjacency matrix

• ghype(default): Generating a ghype model from given xi and omega

• ghype(igraph): Fitting ghype models from an igraph graph

Methods (by generic)

• print(ghype): Print method for ghype object.

Examples

data("adj_karate")
fullmodel <- ghype(graph = adj_karate, directed = FALSE, selfloops = FALSE, unbiased = FALSE)

data('adj_karate')
model <- scm(adj_karate, FALSE, FALSE)
print(model)
#> Call:
#> ghype.matrix(graph = graph, directed = directed, selfloops = selfloops, 
#>     unbiased = TRUE, regular = FALSE)
#> ghype undirected , no selfloops 
#> 34 vertices, 231 edges 
#> Loglikelihood:
#> -434.5449
#> df: 34