This course wraps up the series of methods courses. We look at modelling from a birds-eye view, introducing advanced concepts and revisiting methods introduced in earlier courses from a comprehensive Bayesian perspective. We introduce causal reasoning using directed acyclic graphs (DAGs), mixture models, Gaussian processes, learn to deal with measurement error and missing data; and we revisit regression modelling, generalized linear models, multilevel modelling, Markov chain Monte Carlo sampling, learning to implement them using probabilistic programming.
The course is based on the textbook Statistical Rethinking (2nd edition, 2020) by Richard McElreath1. Please get a copy. The book's homepage contains lots of additional resources. In particular, please install the R package rethinking. Slides and recordings of the author's current course and of previous versions from 2019 and 2022 are also available.
1McElreath, R. (2020). Statistical Rethinking: A Bayesian Course with Examples in R and Stan (2nd ed.). Chapman and Hall/CRC. doi:10.1201/9780429029608
| Course week | Week of year | Topics and readings |
|---|---|---|
| 1 | 6 | Statistical models (chapters 1,2) |
| 2 | 7 | Distributions and sampling (chapters 2,3) |
| 3 | 8 | Gaussian models and linear regression (chapter 4) |
| 4 | 10 | Several predictors, directed acyclic graphs (chapters 5) |
| 5 | 11 | Causal inference (chapter 6) |
| 6 | 13 | Model comparison (chapter 7) |
| 7 | 15 | Interactions (chapter 8) |
| 8 | 16 | Markov chain Monte Carlo, maximum entropy (chapters 9, 10) |
| 9 | 17 | Generalized linear models (chapters 11) |
| 10 | 18 | Mixture models, ordered categorical outcomes/predictors (chapter 12) |
| 11a | 19 | Multilevel models (chapter 13) [only lecture] |
| 11b | 21 | Multilevel models (chapter 13) [only class] |
We will be using a so-called 'flipped classroom' in this course. This means that you are generally expected to have read the literature and watched the vidoes before coming to the lecture. The purpose of the lecture then is for you to ask questions (the more, the better!) so that we can go over the material again together, deepening and broadening our understanding of it.
In order for us to stay in contact, ask questions, and have discussions, there is a Slack workspace dedicated to this course. You will receive an invite link via Brightspace.
-
Portfolio consisting of 3 assignments
-
Each assignment will require you to create an R Markdown notebook consisting of a mix of text and code.
-
Due
- End of week 10 (Sunday 12 March, 23:59)
- End of week 15 (Sunday 16 April, 23:59)
- End of week 18 (Sunday 7 May, 23:59)
You will receive a (short) feedback message from us on your portfolio assignments that you can use for improvements before finalizing your hand-ins.