Bayesian Network
================
.. figure:: _static/BayesNet.png
:align: center
:alt: a Bayesian network (alarm)
The Bayesian Network is the main object of pyAgrum. A Bayesian network is a
probabilistic graphical model. It represents a joint distribution over a set of
random variables. In pyAgrum, the variables are (for now) only discrete.
A Bayesian network uses a directed acyclic graph (DAG) to represent conditional
indepencies in the joint distribution. These conditional indepencies allow to
factorize the joint distribution, thereby allowing to compactly represent very
large ones. Moreover, inference algorithms can also use this graph to speed up
the computations. Finally, the Bayesian networks can be learnt from data.
.. toctree::
:maxdepth: 3
BNModel
BNTools
BNInference
BNLearning