Overview

Statistical Rethinking is an excellent book for applied Bayesian data analysis. The accompanying codes for the book are written in R and Stan. They are then ported to Python language using PyMC3. Recently, Pyro emerges as a scalable and flexible Bayesian modeling tool (see its tutorial page), so to attract statisticians to this new library, I decided to make a Pyronic version for the codes in this repository. Inspired by the PyMC3onic version, I keep the codes in this repository as close as possible to the original codes in the book.

To say a bit more about Pyro, it is a universal probabilistic programming language which is built on top of PyTorch, a very popular platform for deep learning. If you are familiar with numpy, the transition from numpy.array to torch.tensor is rather straightforward (as demonstrated in this tutorial).

Contents

• Preface

• Chapter 1. The Golem of Prague

• Chapter 2. Small Worlds and Large Worlds

• Chapter 3. Sampling the Imaginary

• Chapter 4. Linear Models

• Chapter 5. Multivariate Linear Models

• Chapter 6. Overfitting, Regularization, and Information Criteria

• Chapter 7. Interactions

• Chapter 8. Markov Chain Monte Carlo

• Chapter 9. Big Entropy and the Generalized Linear Model

• Chapter 10. Counting and Classification

• Chapter 15. Horoscopes

Supplements

• rethinking.py

• data

• source

Installation

pip install jupyter pandas pyro-ppl seaborn torch