Login
Intuitive Bayes/Practical MCMC
Buy now

  • $395 or 4 monthly payments of $105

Practical MCMC

  • Course
  • 54 Lessons

Markov Chain Monte Carlo is one of the most utilized tools for modern Bayesian practitioners. It's an algorithm which enables Inference in many settings and models. In this course, we'll provide an in-depth understanding of the algorithm that makes it all possible.
Buy now

Contents

Introduction

Get an overview of MCMC, and learn why it's so relevant for applied data scientists and statisticians. Get familiar with our teaching style where we teach intuitive concepts, filled with practical examples and hands-on code.

Welcome to the MCMC Course
Preview
What do the MC and MC in MCMC stand for?
Preview
What MCMC Enables
Preview
The reality of working with MCMC
Preview
Who's this course for?
Preview
What we'll cover
Preview
Lesson Feedback
Lesson Exercises

Resources Library

Instructions on how access the various resources provided in this course

Intuitive Bayes Discourse Community Invite
Github Repository and Code Access
Environment Installation with Anaconda
Preview
Optional Orientation: Github
Preview
Optional Orientation: Discourse
Preview
Optional Orientation: Podia
Preview

Investigating Inference

Learn the different ways to perform inference, from basic conjugate methods to grid search and their limitations. Then you'll understand why MCMC is a go to tool for today's practitioners.

Introduction
Basic Bayes
Conjugate Models
Grid Search
Lesson Recap
Lesson Feedback
Lesson Exercises
Lesson References

Markov Chain Monte Carlo Deep Dive

Build your own sampler, before moving onto more modern variants such as Hamiltonian Monte Carlo. Learn how MCMC works, and where it sometimes doesn't.
Introduction
MCMC Deep Dive
Introducing Metropolis Hastings
Introducing Hamiltonian Monte Carlo
MCMC in Practice
Lesson Recap
Lesson Feedback
Lesson Exercises
Lesson References

The MCMC Practioners Toolbox

MCMC is paired with complementary tools such as numerical and visual diagnostics. Learn what these tools are, when to use them, and how to interpret their outputs so you can be confident of the results.
Introduction
Diagnostics intuitions
Trace Plots
Rank Plots
R Hat
Autocorrelation and effective size
Divergences
Diagnostics in an end-to-end Bayesian workflow
Lesson Recap
Lesson Feedback
Lesson Exercises
Lesson References

Not So Random Topics

Practical tips for working with MCMC, such as model reparameterization, sampler tuning, and setting priors to get the best results from your sampler.

Introduction
Practical HMC Tuning
(Hierarchical) Reparamaterization
Changing the data
A Cornucopia of Samplers
Monte Carlo Standard Error
Lesson Recap
Lesson Feedback
Lesson Exercises
Lesson References

Final Notes

Congratulations!