Next steps in applied Bayesian regression modeling

Dates

27-31 March 2023

 

To foster international participation, this course will be held online

 

Course overview

Driven by technical innovations and an enthusiastic scientific community, Bayesian data analysis is blooming and booming. While such progress is generally thrilling, it can also become disorienting for those who want to go beyond the very basics: there are so many interesting topics and cool tools that it becomes difficult to see what is useful for the applications that you care about.

This course provides an overview over a number of topics and tools that are not (usually) covered by introductory courses, but that will boost your understanding, productivity and confidence when applying Bayesian regression modeling. The purpose of this course is not exhaustive depth, but guided overview. The goal is to provide enough information (focusing on the conceptual understanding, not mathematical detail) alongside practical examples, to put all participants in a position to learn something new and useful about topics they were already familiar with, to unlock new areas that they heard of but wanted to learn more about, and also to raise awareness for useful ideas and tools that weren't even on the radar of awareness before.

The course also contains interactive practice sessions during which participants can bring up their own concrete use cases and question sets.

Course Prerequisites

The course is not intended as a first introduction to regression modeling or Bayesian data analysis.
It assumes basic familiarity with R, at least simple linear regression, and a rudimentary understanding of Bayesian data analysis (priors, likelihoods, posteriors).
The course will also make use of the tidyverse.
These topics are covered, for example, in this web book, in particular in chapters 2-6 (R and tidyverse), chapters 7-8 (Bayesian foundations), and chapters 12-13 (simple regression).

That being said, the purpose of this course is for everyone to extend their "Bayesian regression comfort zone" in some way or another. So, even if not all of the content covered in the mentioned chapters is familiar to you, this course is still for you.

How to prepare for the course

Before class, please install R 4.1 or higher from the Comprehensive R Archive Network (CRAN). Choose the appropriate "Download R" link depending on your operating system and follow the instructions for downloading and installing R. If you already have R installed, please check that it is the current version (you can check what the 'latest release' of R is by going to CRAN and then compare this with the version shown when you start R). If you do not have the latest version installed, please update.

Although not strictly necessary, it will be useful to also install an integrated development environment (IDE) for R. A popular choice is RStudio. So, unless you already have a different setup, please download and install RStudio.

Please also install the following R packages (ideally: latest versions available):
- tidyverse
- brms

Program

Daily schedule:
2 - 6  pm (Berlin time):
live lectures,  R walkthrough and introduction to / review of the practicals


2/3 additional hours: self-guided practicals using annotated scripts, with live remote support via Slack

 

Monday– Classes from 2-6 PM Berlin time

1. Introduction, practicalities, course overview, recap
2. Understanding Bayesian regression models through visualization


Tuesday– Classes from 2-6 PM Berlin time

1. Unpacking HMC to cope with diagnostics, warnings & error messages
2. How to deal with missing or uncertain data


Wednesday– Classes from 2-6 PM Berlin time

 

1. Bayesian model comparison in practice (from cross-validation to Bayes factors and back)
2. Mastering categorical predictors (encoding schemes and their interpretation)


 
Thursday– Classes from 2-6 PM Berlin time

1. Distributional and mixture models
2. Less standard link & likelihood functions (e.g., ordered logit, Poisson, beta regression ...)


Friday– Classes from 2-6 PM Berlin time

 

1. Causal inference
2. [topics of interest to the audience / general discussion]

 

 

 

 


Cost overview

Package 1

 

480 €

 


Should you have any further questions, please send an email to info@physalia-courses.org

Cancellation Policy:

 

> 30  days before the start date = 30% cancellation fee

< 30 days before the start date= No Refund.

 

Physalia-courses cannot be held responsible for any travel fees, accommodation or other expenses incurred to you as a result of the cancellation.