Instructor: Adam Loy
Location: CMC 319
Time: 4a
Office hours: Mon 2-3, Tue 2-3, Wed 9:30-10:30, Fri 9:30-10:30, and by appointment
Jump to: Daily schedule
The required textbook is Bayesian Statistical Methods by Reich and Ghosh (2019, CRC Press) both readings and problems will be assigned from this text, so please obtain a copy.
I will assign supplemental readings throughout the term and will post links to them on this page.
One major supplement to the course will be the textbook Bayesian Ideas and Data Analysis, which is freely available to Carls on ProQuest Ebook Central
We will use R/RStudio as our computational engine. Please install the current version of R and RStudio at the beginning of the term if you plan to use a local version.
-
Download R: https://cran.r-project.org/
-
Download RStudio desktop: https://www.rstudio.com/products/rstudio/download/#download
The mirage server (https://mirage.mathcs.carleton.edu) is also available for your use, but can only be accessesed on campus of via a VPN.
We will use JAGS (Just Another Gibbs Sampler) “automate” some of the posterior sampling via MCMC this term. Please install JAGS using prior to installing the R packages listed below.
Link to the latest version: https://sourceforge.net/projects/mcmc-jags/files/latest/download
We will use numerous R packages throughout this course. They are all installed on the mirage R Studio server. If you are working on a local install, then please run the code chunk below to install all of the packages. I recommend doing this at the beginning of the course to avoid last minute installation issues preventing you from completing assignments.
install.packages(c("mvtnorm", "loo", "coda", "rjags"), dependencies = TRUE)
I will post homework assignments and their solutions here. Check the folders at the top
Links to any slides and handouts are in the calendar section below. Check the docs folder above for the .Rmd files containing the R code used to generate slides, etc.
Check the exams folder for a study guide (list of topics) and practice problems.
Homework
-
Individual assignments will be due most Tuesdays and Fridays by 4:00 p.m.
-
Team assignments will be assigned roughly every two weeks.
Exam 1: Wednesday, October 16
Case study 1: Wednesday, October 23 by 4 pm
Case study 2: Friday, November 8 by 4 pm
Exam 2: Wednesday, November 13
-
Team assignment questionnaire: 4 November by 4:00 p.m.
-
Proposal and data: 11 November by 4:00 p.m.
-
Final submission: 25 November by 3:00 p.m.
-
Email me the report as a PDF
-
Email me (or share via Dropbox/Google Drive) the supplemental materials: data, code
-
BSM = Bayesian Statistical Methods
BIDA = Bayesian Ideas and Data Analysis
| Date | Reading | Topic | Materials |
|---|---|---|---|
| 16-Sep | BSM 1.1-1.2; BIDA 2.1 |
Welcome and probability review | Slides |
| 18-Sep | BSM 1.3; BIDA 2.2-2.3 |
Comparing paradigms, updating prior belief, discrete priors | Slides |
| 20-Sep | 1.4.1-1.4.2; BIDA 2.4 |
Continuous priors, posterior analysis | Slides |
| 23-Sep | BSM 1.5; BIDA 3.1 |
Posterior analysis, Monte Carlo simulation, grid approximation | Slides |
| 25-Sep | BSM 1.5 | Bayesian prediction, model checking, conjugate analysis | birthorder.csv |
| 27-Sep | BSM 2.1.1-2.1.2; BIDA 2.2-2.3 |
Selecting conjugate priors | Slides |
| 30-Sep | BSM 2.1.3-2.1.5, 2.1.8 | More conjugate priors, mixtures of conjugate priors | Slides |
| 2-Oct | BSM 2.2-2.3.1; BIDA 4.6 |
Natural conjugate priors, nonconjugate priors, uninformative priors | Slides |
| 4-Oct | BSM 1.4.3; BIDA 4.7 |
Jeffreys’ prior, intro to multiparameter models | Slides |
| 7-Oct | BIDA 5.2-5.2.3 | Multiparameter models: One-sample models | Slides |
| 9-Oct | BIDA 5.1.3, 5.2.5, 5.3.4 | Multiparameter models: Two-sample models | |
| 11-Oct | BSM 3.1 | Deterministic Bayesian computing | Slides |
| 14-Oct | BSM 3.1 | Deterministic Bayesian computing (continued) | Slides; Bayesian CLT supplement |
| 16-Oct | Review 1.1-3.1 | Exam 1 | Study guide; Practice problems; Distributions |
| 18-Oct | Supplemental reading | Introduction to Markov chains | Outline of notes; Example solution |
| 21-Oct | No class - Midterm break | ||
| 23-Oct | BSM 3.2-3.2.1 | Gibbs sampling | Slides; Handout |
| 25-Oct | BSM 3.2.2 | Metropolis sampling | Slides; Gibbs example solution |
| 28-Oct | BSM 3.2.2 | Metropolis-Hastings sampling | Slides |
| 30-Oct | BSM 3.3 | Intro to JAGS + convergence issues | Slides; JAGS user manual; JAGS example .Rmd (right click to download); Example solutions |
| 1-Nov | BSM 3.4 | MCMC diagnostics | MCMC diagnostic slides |
| 4-Nov | BSM 4.1-4.2.2 | Bayesian linear regression | Fitting BLR in R/JAGS (.Rmd) |
| 6-Nov | BSM 4.2.2-4.2.3 BIDA 9.4 |
Informative + regularizing priors | Slides Video part I Video part II |
| 8-Nov | BSM 5.1, 5.5 | Model comparisons | Slides |
| 11-Nov | Wrap up and review | ||
| 13-Nov | Review | Exam 2 | Study guide; Practice problems; Practice solutions |
| 15-Nov | BSM 4.3 | Generalized linear models | Slides; UC Berkley example |
| 18-Nov | BSM 5.6 | Posterior predictive checks | Slides |
| 20-Nov | Project work day |