checkdegeneracy: Check an cergm object for model degeneracy

In schmid86/cERGM: A package to fit citation Exponential Random Graph Models

Description

Calculates observed statistics and percentiles of statstics of simulated networks. If the observed statistc differs significantly from the simulated statistics, then this might indicate model degeneracy. This function depends on the ergm-package.

Usage

degeneracycheck(object, ...)

Arguments

object

An cergm object

Author(s)

Christian S. Schmid <[e-mail]>

References

Hunter D, Handcock M, Butts C, Goodreau S, Morris M (2008). ergm: A Package to Fit, Simulate and Diagnose Exponential-Family Models for Networks. Journal of Statistical Software, 24(3), 1-29.

Schmid C, Chen T, Desmarais B (2020). Generative Dynamics of Supreme Court Citations: Analysis with a New Statistical Network Model.

Examples

# load Supreme Court Citation Network from 1936-1941
data("scc_1936_1941")

# get vertex attribute "Term". Indicates the Term of each node
terms <- get.vertex.attribute(scc_1936_1941, "Term")

# create a matrix with the sender's term in each row
terms.matrix <- matrix(terms,length(terms),length(terms),byrow=F)

# fix all dyads that can potentially be created in 1941 as 1
unfixed.dyads <- 1*(terms.matrix == 1941)


# alternatively one can also add an unfixed vector that indicates
# which nodes can create edges
unfixed.vector<-  1*(terms==1941)

# test cergm-function
model <- cergm(scc_1936_1941~ edges+ difftransties("Term")+ nodeicov("NumberJusticesPro"),
            not.fixed=unfixed.dyads, estimate="MPLE", init.method="SAN")

checkdegeneracy(model)

schmid86/cERGM documentation built on Sept. 10, 2021, 6:20 p.m.