Citation Context

...hods developed for static networks and parametrize the transition between graphs. Examples include time-varying latent space methods [16], latent feature models [17], time-lagged logistic regressions =-=[18]-=-, and temporal exponential random graph models (ERGMs) [19, 20] using specifications analogous to static ERGMs. A separate line of work assumes the observed networks are outcomes from a latent continu...

Citation Context

...itioned on covariates such as the network users’ group memberships or seasonal effects (e.g. the day of the week). Such models can be easily learned through logistic regression and were considered by =-=[3]-=-, who termed them Dynamic Network Regression (DNR) families. However, the conditional independence assumptions of DNR can be too restrictive for some applications, especially when few covariates are a...