link_significance: Estimate statistical deviations from ghype model

In gi0na/ghypernet: Fit and Simulate Generalised Hypergeometric Ensembles of Graphs

Estimate statistical deviations from ghype model

Description

linkSignificance allows to estimate the statistical deviations of an observed graph from a ghype model.

Usage

linkSignificance(
  graph,
  model,
  under = FALSE,
  log.p = FALSE,
  binomial.approximation = FALSE,
  give_pvals = FALSE
)

link_significance(
  graph,
  model,
  under = FALSE,
  log.p = FALSE,
  binomial.approximation = FALSE,
  give_pvals = TRUE
)

Arguments

graph	an adjacency matrix or a igraph object.
model	a ghype model
under	boolean, estimate under-represented deviations? Default FALSE: i.e. returns over representation
log.p	boolean, return log values of probabilities
binomial.approximation	boolean, force binomial? default FALSE
give_pvals	boolean, return p-values for both under and over significance? when FALSE, it returns probabilty of observing stricly more (or less) edges than in graph. When TRUE returns probability of observing exactly as many edges or more (less) than in graph, like a standard pvalue.

Value

matrix of probabilities with same size as adjacency matrix.

Examples

data("adj_karate")
fullmodel <- ghype(graph = adj_karate, directed = FALSE, selfloops = FALSE)
link_significance(graph = adj_karate, model = fullmodel, under=FALSE)

gi0na/ghypernet documentation built on April 13, 2024, 2:33 a.m.