ego_tergm: Estimation of ego-Temporal Exponential Random Graph Model...

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

Description

This function estimates an ego-TERGM on a longitudinally observed network. Currently the function does not support comparisons of whole networks.

Usage

ego_tergm(net = NULL, form = NULL, core_size = 1, min_size = 5,
  roles = 3, add_drop = TRUE, directed = TRUE,
  edge_covariates = FALSE, seed = 12345, R = 10, forking = FALSE,
  ncpus = 1, steps = 50, tol = 1e-06)

Arguments

net	The longitudinally observed network that an ego-TERGM will be fit on. Must be presented as a list of networks. 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.
add_drop	Do nodes drop out of the network or enter it? If so, specify as the default TRUE.
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. These should be stored as network attributes.
seed	The seed set to replicate analysis for pseudorandom number generator.
R	The number of bootstrap replications that should be used for the estimation of a bootstrapped MPLE estimated TERGM for model initialization. Defaults to 10.
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 and input values, including net (original networks), lambda (the probability of assignments), group.theta (the roles by terms cluster centroids), EE.BIC (the Salter-Townshend and Murphy BIC cross-sectional BIC), TS.BIC (the Campbell BIC penalizing for time-steps), role_assignments (a data frame of the most likely assignments), reduced_networks (A list of the networks with excluded egos), ego_nets (a list of ego-networks), and ego_nets_used (N x T matrix of logicals here TRUE refers to ego-networks kept).

References

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

Leifeld, Philip, Skyler J. Cranmer and Bruce A. Desmarais (2017): Temporal Exponential Random Graph Models with btergm: Estimation and Bootstrap Confidence Intervals. Journal of Statistical Software 83(6): 1-36. http://dx.doi.org/10.18637/jss.v083.i06

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"))

ego_tergm_fit <- ego_tergm(net = net,
                          form = c("edges", "mutual", "triangle",
                                   "nodeicov('idegsqrt')", "nodeocov('odegsqrt')",
                                   "nodematch('sex')"),
                          core_size = 1,
                          min_size = 5,
                          roles = 3,
                          add_drop = TRUE,
                          directed = TRUE,
                          edge_covariates = FALSE,
                          seed = 12345,
                          R = 10,
                          forking = FALSE,
                          ncpus = 1,
                          steps = 50,
                          tol = 1e-06)

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