Course Overview
Bayesian methods are increasingly important in both industry and academia. This is an upper level undergraduate-course for students registered in STA 360, and a graduate-level course for students registered in STA 602. This is an asynchronous online course that introduces students to the basics of Bayesian inference and provides students with the tools needed to fit Bayesian models. Asynchronous here means that there will be no standard meeting time, so that students are able to participate according to their own schedule. That said, the course is designed to align with the standard six-week Term II schedule of the summer school calendar. Therefore, there will be set deadlines to ensure that all course materials and assessments are completed in six weeks.
In this course, you will learn the importance of Bayesian methods and inference. You will be introduced to Bayesian theory, with particular emphasis on conceptual foundations as well as implementation and model fitting. You will learn the essential distinctions between classical and Bayesian methods and become familiar with the origins of Bayesian inference. You will also learn about conjugate families of distributions and why they are very convenient, and how to conduct Bayesian inference with intractable posterior distributions, when you do not have conjugate distributions.
Although this course emphasizes the mathematical theory behind Bayesian inference, data analysis and interpretation of results are also important components. Students who wish to explore the mathematical theory in more detail than what is covered in class are welcome to engage with and request further reading materials from the instructor. Also, all students must have the theoretical background covered in the prerequisites to be able to keep up with and understand the materials.
Learning Objectives
By the end of this course, students should be able to
- Understand the basics of Bayesian inference, that is, be able to define likelihood functions, prior distributions, posterior distributions, prior predictive distributions and posterior predictive distributions.
- Derive posterior distributions, prior predictive distributions and posterior predictive distributions, for common likelihood-prior combinations of distributions.
- Interpret the results of fitted models and conduct checks to ascertain that the models have converged.
- Use the Bayesian methods and models covered in class to analyze real data sets.
- Assess the adequacy of Bayesian models to any given data and make a decision on what to do in cases when certain models are not appropriate for a given data set.
Course Info
Playposit
To gain access to the pre-recorded lecture videos, you will have to enroll in Playposit via Duke. There are participation quizzes embedded within the videos. These quizzes make up 5% of your final grade (see: course policies) so take them seriously. To join the class on Playposit, you need create a new account as a student here. Next, you will use the class link, which I will send out via email, to join the class site.
Zoom meetings
The easiest way for you to join the different Zoom meetings is to log in to Sakai, go to the "Zoom meetings" tab, and click "Upcoming Meetings". For the recordings (for lab and discussion sessions), also log in to Sakai, go to the "Zoom meetings" tab, and click "Cloud Recordings". Those will be available few minutes after the sessions.
Meeting Times
There will be no fully synchronous meetings. There will be pre-recorded lecture videos for each topic. There will be Zoom discussion sessions every day (see below for times), but those will be recorded and shared with all students. Students are strongly encouraged to attend the live sessions and ask questions but students who prefer not to can watch the recordings.
Monday - Friday: 10:00am - 11:00am
Zoom Meeting ID: See Sakai.
Online Labs
These are live optional lab sessions on Zoom for students who wish to attend. Recordings will be made available afterwards for students who prefer not to.
Mondays:
2:45pm - 4:00pm
Zoom Meeting ID: See Sakai
Wednesdays:
11:30am - 12:45pm
Zoom Meeting ID: See Sakai
Teaching Team and Office Hours
| Instructor | Dr. Olanrewaju Michael Akande | Mondays: 6:00pm - 7:00pm Wednesdays: 9:00am - 10:00am Thursdays: 12:00pm - 1:00pm |
Zoom Meeting ID: See Sakai Zoom Meeting ID: See Sakai Zoom Meeting ID: See Sakai |
|
| TA | Christine Shen | Mondays: 4:00pm - 5:00pm Tuesdays: 8:00pm - 10:00pm |
Zoom Meeting ID: See Sakai Zoom Meeting ID: See Sakai |
|
| Bo Liu | Fridays: 3:00pm - 5:00pm | Zoom Meeting ID: See Sakai |
Texts
| A First Course in Bayesian Statistical Methods | Peter D. Hoff, 2009, New York: Springer. | Required (available online from Duke library) |
| Bayesian Data Analysis (Third Edition) | Andrew Gelman, John Carlin, Hal Stern, David Dunson, Aki Vehtari, and Donald Rubin. | Optional |
Materials
Lecture notes and slides, lab exercises and assigned readings will be posted on the course website, while discussion and lab videos will be posted on Sakai (Zoom via Sakai). Lecture videos will be posted on Playposit. White boards will also be used frequently in the lecture videos, so please pay special attention to those. Finally, we will closely follow the main textbook so students should make sure to always read the corresponding textbook chapters in the assigned readings.
Important Dates
| Mon, June 29 | Classes begin |
| Wed, July 1 | Drop/Add for Term II ends |
| Fri, July 3 | Independence Day holiday observed |
| Tue, July 7 | Quiz I day (tentative) |
| Fri, July 17 | Midterm exam day (tentative) |
| Mon, July 27 | Last day to withdraw with W |
| Wed, July 29 | Quiz II day (tentative) |
| Thurs, August 6 | Classes end |
| Fri, August 7 - Sun, August 9 | Final exam period |
Green Classroom
This course has achieved Duke’s Green Classroom Certification. The certification indicates that the faculty member teaching this course has taken significant steps to green the delivery of this course. Your faculty member has completed a checklist indicating their common practices in areas of this course that have an environmental impact, such as paper and energy consumption. Some common practices implemented by faculty to reduce the environmental impact of their course include allowing electronic submission of assignments, providing online readings and turning off lights and electronics in the classroom when they are not in use. The eco-friendly aspects of course delivery may vary by faculty, by course and throughout the semester. Learn more at http://sustainability.duke.edu/action/certifications/classroom/index.php.
Acknowledgement
This web page contains materials such as lecture slides, homework assignments, and datasets developed or adapted by Dr. Alexander Volfovsky, Dr. David B. Dunson and Dr. Rebecca Carter Steorts.