Hello Mr. Dustin
My name is Robson Fernandes. I am a master student in Mathematics, Statistics and Computing at the University of São Paulo, USP.
I have researched several frameworks to work with probabilistic models, in particular Gaussian Bayesian networks.
I would like to know if the Edward framework, developed by you has support for the assembly of this type of network “Gaussian Bayesian networks”.
I have used the bnlearn library in R, but would love to find a full python alternative.
Thank you very much, and congratulations for your projects!
I understand “Bayesian networks” as the classical name for “directed graphical models”. Are “Gaussian Bayesian networks” Bayesian networks with only Gaussian distributed nodes? Also, how is inference typically performed in these models: MAP, variational inference, etc.?
There are plentiful examples in the tutorials and
examples/ repo using (mostly) normal random variables.
Exactly, they are directed graphical models where the nodes have Gaussian distribution.
I’m learning to work with the framework yet, sorry, but I’m a beginner.
I would like to get the prediction of a Gaussian value.
I have a “directed graphical models” with this structure.
I´d like inference the Y1 node and get a value predict from test data set.
The inference will be done with “variational inference”
Where I can found a example about this?