Conditional Graphical Model Family

A Conditional Graphical Model Family is a graphical statistical model family that restricts itself to representing conditional probabilities as conditional probability networks.

• AKA: Conditional Graphical Metamodel, Conditional Graphical Models, Conditional Probabilistic Graphical Models.
• Context:
  • It can be associated with a Conditional Probabilistic Network Instance.
  • It can range from being an Undirected Conditional Graphical Metamodel to being a Directed Conditional Graphical Metamodel.
  • Every path from a node in X to a node in Y must pass through at least one node in Z.
• Example(s):
  • Conditional Random Fields Metamodel.
  • Maximum-Margin Markov Networks.
  • …
• Counter-Example(s):
  • a Conditional Probability Network.
  • a Non-Conditional Probability Graphical Metamodel.
• See: Probabilistic Graphical Metamodel.

---

---

References

2007

• (Perez-Cruz et al., 2007) ⇒ Fernando Perez-Cruz, Massimiliano Pontil, and Zoubin Ghahramani. (2007). “Conditional Graphical Models.” In: (Bakir et al., 2007)

1998

• Kevin Murphy. (1998). “A Brief Introduction to Graphical Models and Bayesian Networks.".
  • http://www.cs.ubc.ca/~murphyk/Bayes/bnintro.html

---