Background:

Probablistic Graphical Models representation in Python

Probablistic Graphical Models (PGMs) use a graph-based representation to compactly encode the joint distribution. PGMs use elements of graph theory and probability theory. There are two classes of PGMs. directed acyclic graphical models are called Bayesian Networks (BNs) and undirected graphical models are Markov networks.

The code in this repository is based off the Matlab code from Daphne Koller's Coursera class on PGMs https://www.coursera.org/course/pgm For the assignments in the class students were given starter Matlab code and had to implement various algorithms, etc.

While I grew to like Matlab, I needed to implement the inference and represenation of PGMs in Python for a personal project of mine. So I went down the rabbit hole. The code here depends on NumPy and uses Python v2.7.

Amit Indap