PGMs
PGMs are a formalism used to represent probabilistic relationships between random variables. They
are appropriate to model problems with uncertainty.
Usually, a PGM comprises some sort of graphical structure that captures the independence/dependence relationships
between the variables, and parametric tables are represent the
strength of these relationships.
Our research on PGMs has been mainly oriented in the following directions:
Copulas
Markov networks
Regularization methods
EDAs