Dynamic PGN (DPGN)
Mohammed M. Ettouney · 2025
The links between the variables in the CC- and CI worlds are either directional, as in Bayesian Networks (BNs) or bi-directional, as in Markov networks (MNs). The cause / effect directionality of the former or the feedback bidirectionality of the latter implies time lag between the occurrences / materializations of the variables. Sometimes, it is sufficient to ignore the time lags between variables and model the network in a snapshot model. In many other situations the time lags can be long enough, e.g., sea level rise (SLR), to require the consideration of consecutive snapshots at different instances of time. These multi-snapshot models, dynamic probabilistic graph networks (DPGNs), are of interest.