search

dkesada / dbnR

Gaussian dynamic Bayesian networks structure learning and inference based on the bnlearn package
GNU General Public License v3.0
44 stars 10 forks source link

arrow point from t_2 to t_1 and t_0 #9

Closed zh-zhang1984 closed 2 years ago

zh-zhang1984 commented 2 years ago

Hi I have a question after learning the tutorial. I found the plot drawing by dbnR::plot_dynamic_network(fit) The arrow direction is from _t_2 to t_1 and then to t_0; I am confusing this because the causal direction can only from t_0 to t_1 and t_2, events happened early can affect latter events, but not in reversed order.

dkesada commented 2 years ago

Hi @zh-zhang1984 We changed the naming convention a little in dbnR and made t_0 the most recent time slice instead of the oldest one. In that example, the oldest events happen at t_2, and that affects the more recent t_1 and t_0. Given that we allow any Markovian order to be used for the DBNs, it is more convenient for inference purposes to always have t_0 as the most recent time slice. You can think of t_0 as the most recent event.

zh-zhang1984 commented 2 years ago

Thank you for the clarification!