Cegpy (/segpaɪ/) is a Python package for working with Chain Event Graphs. It supports learning the graphical structure of a Chain Event Graph from data, encoding of parametric and structural priors, estimating its parameters, and performing inference.
It is built on top of the Python network modelling package NetworkX.
Documentation is hosted on read the docs.
We have also written a paper to explain the statistical methods and algorithms included in the package; ARXIV - cegpy: Modelling with Chain Event Graphs in Python.
If you'd like to get started using the packages, the best place to start is the quick-start documentation.
Use cases have been demonstrated in a binder.
The following image is an example of a chain event graph that is produced by this package:
Thanks goes to these wonderful people (emoji key):
Aditi Shenvi 💻 |
Gareth Walley 💻 📖 🎨 |
Kasia Kobalczyk 💻 🐛 |
Peter Strong 💻 🐛 💡 |
This project follows the all-contributors specification. Contributions of any kind welcome!