search

matheusfacure / python-causality-handbook

Causal Inference for the Brave and True. A light-hearted yet rigorous approach to learning about impact estimation and causality.
https://matheusfacure.github.io/python-causality-handbook/landing-page.html
MIT License
2.72k stars 472 forks source link

Feature Request - BART #267

Open bstuddard opened 2 years ago

bstuddard commented 2 years ago

I really enjoyed the various chapters, writing style, and code examples. Any thoughts on doing a chapter on Bayesian Additive Regression Trees and their use as a method in causal inference?

luruismert commented 1 year ago

There is another python based book project here on Github which helped me to understand the concept of BART models. It is about bayesian modeling and includes an chapter on BART models. However, I think the code is not up to date with the PyMC library. The latest published version of 'An Introduction to statistical learning with Applications in Python', which was published just 2 month ago, also includes an chapter on BART models and can be used as supplementary material for the first mentioned source. The issue with this book (in my opinion) is the lack of generelizable code for implementing a BART model with an up to date flexible library.

PyMC seems to be the most flexible and up to date option for an implementation (look here for a quick introduction to the model and its implementation: https://www.pymc.io/projects/examples/en/latest/case_studies/BART_introduction.html).

I think that adding a chapter that helps to understand how we can draw inference from BART models and provide intuitive up to date code on the implementation would be a great addition to the project.