"btergm" and "mtergm"btergm objects result from the estimation of a bootstrapped TERGM via the btergm function in the xergm package. btergm objects contain the coefficients, the bootstrapping samples of the coefficients, the number of replications, the number of observations, the number of time steps, the original formula, and the response, effects and weights objects that were fed into the glm call for estimating the model. mtergm objects result from MCMC-MLE-based estimation of a TERGM via the mtergm function. They contain the coefficients, standard errors, and p-values, among other details.
# S4 method for btergm
summary(object, level = 0.95, type = "perc",
invlogit = FALSE, ...)# S4 method for mtergm
summary(object, ...)
# S4 method for btergm
show(object)
# S4 method for mtergm
show(object)
# S4 method for btergm
nobs(object)
# S4 method for mtergm
nobs(object)
# S4 method for btergm
coef(object, invlogit = FALSE, ...)
# S4 method for mtergm
coef(object, invlogit = FALSE, ...)
# S4 method for btergm
confint(object, parm, level = 0.95, type = "perc",
invlogit = FALSE, ...)
btergm.se(object, print = FALSE)
timesteps.btergm(object)
timesteps.mtergm(object)
A btergm object.
The significance level for computation of the confidence intervals. The default is 0.95 (that is, an alpha value of 0.05). Other common values include 0.999, 0.99, 0.9, and 0.5.
Parameters (specified by integer position or character string).
Type of confidence interval, e.g., basic bootstrap interval (type = "basic"), percentile-based interval (type = "perc", which is the default option), or bias-adjusted and accelerated confidence interval (type = "bca"). All options from the type argument of the boot.ci function in the boot package can be used to generate confidence intervals.
Apply inverse logit transformation to the estimates and/or confidence intervals? That is, 1 / (1 + exp(-x)), where x is the respective value.
Should the formatted coefficient table be printed to the R console along with significance stars (print = TRUE), or should the plain coefficient matrix be returned (print = FALSE)?
Further arguments to be handed over to subroutines.
coef:Object of class "numeric". The coefficients.
bootsamp:Object of class "matrix". The bootstrapping sample.
R:Object of class "numeric". Number of replications.
nobs:Object of class "numeric". Number of observations.
time.steps:Object of class "numeric". Number of time steps.
formula:Object of class "formula". The original model formula (without indices for the time steps).
response:Object of class "integer". The response variable.
effects:Object of class "data.frame". The effects that went into the glm call.
weights:Object of class "integer". The weights of the observations.
auto.adjust:Object of class "logical". Indicates whether automatic adjustment of dimensions was done before estimation.
offset:Object of class "logical". Indicates whether an offset matrix with structural zeros was used.
directed:Object of class "logical". Are the dependent networks directed?
bipartite:Object of class "logical". Are the dependent networks bipartite?
Standard errors.
p-values.
Estimate: either MCMC MLE or MPLE.
The log likelihood.
Akaike's Information Criterion (AIC).
The Bayesian Information Criterion (BIC).
Various generic methods are available for btergm objects: The coef and show methods return the coefficients; the summary method gives a model summary. The nobs method returns the number of observations. The confint method returns confidence intervals from the bootstrap replications of btergm objects, and the user can specify the confidence level. The method returns a matrix with three columns: the estimate, the lower bound, and the upper bound of the confidence interval for each model term.
The btergm.se function computes standard errors and p values for btergm objects. It returns a matrix with four columns: the estimate, the standard error, the z value, and the p value for each model term. If the argument print = TRUE is used, the matrix is printed to the R console as a formatted coefficient matrix with significance stars instead. Note that confidence intervals are the preferred way of interpretation for bootstrapped TERGMs; standard errors are only accurate if the bootstrapped data are normally distributed, which is not always the case. Various methods for checking for normality for each model term are available, for example quantile-quantile plots (e.g., qqnorm(x@boot$t[, 1]) for the first model term in the btergm object called x).
The timesteps.btergm function extracts the number of time steps from a btergm object. The number of time steps is the number of networks being modeled on the left-hand side of the model formula.
Some of these functions or methods are also available for mtergm objects.
Cranmer, Skyler J., Tobias Heinrich and Bruce A. Desmarais (2014): Reciprocity and the Structural Determinants of the International Sanctions Network. Social Networks 36(1): 5--22. http://dx.doi.org/10.1016/j.socnet.2013.01.001.
Desmarais, Bruce A. and Skyler J. Cranmer (2012): Statistical Mechanics of Networks: Estimation and Uncertainty. Physica A 391: 1865--1876. http://dx.doi.org/10.1016/j.physa.2011.10.018.
Desmarais, Bruce A. and Skyler J. Cranmer (2010): Consistent Confidence Intervals for Maximum Pseudolikelihood Estimators. Neural Information Processing Systems 2010 Workshop on Computational Social Science and the Wisdom of Crowds.
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.
btergm-package btergm simulate.btergm gofmethods knecht getformula interpret mtergm