ego_ergm: Estimation of ego-Exponential Random Graph Model (ego-ERGM)...

In benjamin-w-campbell/egoTERGM: Estimation of Ego-Temporal Exponential Random Graph Models via Expectation Maximization (EM)

Description

This function estimates an ego-ERGM. Code taken from Salter-Townshend and Murphy (2015)'s replication archive.

Usage

ego_ergm(net = NULL, form = NULL, core_size = 1, min_size = 5,
  roles = 3, directed = TRUE, edge_covariates = FALSE,
  seed = 12345, forking = FALSE, ncpus = 1, steps = 50,
  tol = 1e-06)

Arguments

net	The cross-sectional network that an ego-ERGM will be fit on. Must be presented as a network object. Any vertex attributes should be attached to networks. Currently the function does not support comparisons of whole networks.
form	The formula comprised of ERGM or TERGM terms used to distinguish between clusters assignments. Specified as a vector of comma separated terms. No default.
core_size	The order of alters to include. The default value of one implies only looking at an ego's alters and the connections among them.
min_size	The minimum number of nodes an ego-network must achieve to be included. Defaults to five.
roles	The number of roles that should be fit. Defaults to 3.
directed	Should the longitudinal network be treated as directed? If so, specify as the default TRUE.
edge_covariates	Are edge covariates included in the form term? IF so, specify as TRUE. No default.
seed	The seed set to replicate analysis for pseudorandom number generator.
forking	If parallelization via forking should be used (TRUE) or if no parallel processing should be used (FALSE). Currently, sockets are not supported.
ncpus	The number of CPUs that should should be used for estimation, defaults to 1.
steps	The number of default EM steps that should be taken, defaults to 50.
tol	The difference in parameter estimates between EM iterations to determine if the algorithm has converged. Defaults to 1e-6.

Value

A list of model results, including lambda (the probability of assignments), group.theta (the roles by terms cluster centroids), EE.BIC (the Salter-Townshend and Murphy BIC cross-sectional BIC), role_assignments (a data frame of the most likely assignments), and reduced_networks (network with excluded ego).

References

Box-Steffensmeier, Janet M., Benjamin W. Campbell, Dino P. Christenson, Zachary Navabi. (2018): Role analysis using the ego-ERGM: A Look at environmental interest group coalitions. Social Networks 52: 213-227. https://doi.org/10.1016/j.socnet.2017.08.004

Campbell, Benjamin W. (2018): Inferring Latent Roles in Longitudinal Networks. Political Analysis 26(3): 292-311. https://doi.org/10.1017/pan.2018.20

Salter-Townshend, Michael and Thomas Brendan Murphy. (2015): Role Analysis in Networks using Mixtures of Exponential Random Graph Models. Journal of Computational and Graphical Statistics 24(2): 520-538. https://doi.org/10.1080/10618600.2014.923777

Examples

# Code from xergm.common and their preparation of the Knecht network
library(xergm.common)
set.seed(1)

data("knecht")

for (i in 1:length(friendship)) {
 rownames(friendship[[i]]) <- paste("Student.", 1:nrow(friendship[[i]]), sep="")
 colnames(friendship[[i]]) <- paste("Student.", 1:nrow(friendship[[i]]), sep="")
}
rownames(primary) <- rownames(friendship[[1]])
colnames(primary) <- colnames(friendship[[1]])
sex <- demographics$sex
names(sex) <- rownames(friendship[[1]])
# step 2: imputation of NAs and removal of absent nodes:
friendship <- xergm.common::handleMissings(friendship, na = 10, method = "remove")
friendship <- xergm.common::handleMissings(friendship, na = NA, method = "fillmode")
# step 3: add nodal covariates to the networks
for (i in 1:length(friendship)) {
  s <- xergm.common::adjust(sex, friendship[[i]])
  friendship[[i]] <- network::network(friendship[[i]])
  friendship[[i]] <- network::set.vertex.attribute(friendship[[i]], "sex", s)
  idegsqrt <- sqrt(sna::degree(friendship[[i]], cmode = "indegree"))
  friendship[[i]] <- network::set.vertex.attribute(friendship[[i]],
                                                   "idegsqrt", idegsqrt)
  odegsqrt <- sqrt(sna::degree(friendship[[i]], cmode = "outdegree"))
  friendship[[i]] <- network::set.vertex.attribute(friendship[[i]],
                                                   "odegsqrt", odegsqrt)
}
sapply(friendship, network::network.size)
net <- friendship
rm(list=setdiff(ls(), "net"))

# Reduce down to first time-step
ego_ergm_fit <- ego_ergm(net = net[[1]],
                          form = c("edges", "mutual", "triangle",
                                   "nodeicov('idegsqrt')", "nodeocov('odegsqrt')",
                                   "nodematch('sex')"),
                          core_size = 1,
                          min_size = 5,
                          roles = 3,
                          forking = FALSE,
                          ncpus = 1,
                          directed = TRUE,
                          edge_covariates = FALSE,
                          seed = 12345,
                          steps = 50,
                          tol = 1e-06)

benjamin-w-campbell/egoTERGM documentation built on June 3, 2019, 5:56 p.m.