STATISTICS 530 (2009-2010, Term 2)

BAYESIAN INFERENCE AND DECISION (3 Credits)

Lectures: Monday and Wednesday 9:30-11:00, LSK 301.
Evolving schedule.

Instructor: Paul Gustafson, LSK 326, gustaf at stat dot ubc dot ca.
Office hours are Wed. 11-12 and Fri. 9:30-10:30. (Ordinarily with a graduate course I would try `open-door' office hours, but the current enrolment is way beyond ordinary.)

Prerequisites: Open to any interested graduate students in the Department of Statistics. Graduate students from other departments are welcome, provided they have sufficient statistical and mathematical backgrounds. Such students should consult the instructor about suitability.

Description:: A wide-ranging investigation of the Bayesian approach to statistical inference. Conceptual, computational, practical, and theoretical issues will be discussed.

Website (this page): www.stat.ubc.ca/~gustaf/stat530.html

Textbook: A first course in Bayesian statistical methods by Peter Hoff. Springer, 2009.

The book is freely available in pdf to the UBC community via the library's e-book collection (try this direct link), and the bookstore has copies for those liking old-fashioned books. The author maintains a website with supporting materials.

Other References: There are many Bayesian books out there currently (Bolstad, Carlin and Louis, Congdon, Gelman et. al., Lee, ...)

Other Resources:

Lecture format: I envision two modes of delivery:

1. When we are discussing chapters from the text, I will use informal `whiteboard' presentation, and I hope that for the most part we can have interactive discussion rather than lecturing. It will be expected that you have read the chapter in question in advance (the discussion below on class participation is very relevant to this point).

2. When we are discussing ideas that go beyond the text, I will prepare material more formally, and will likely use slides.

Evaluation: equal weight on (i) class participation, (ii), the first assignment, (iii) the second assignment, (iv) a term paper in lieu of a final exam.

Class participation: The primary issue here is that for lectures devoted to discussion of chapters from the text, you read the chapter in advance. As some evidence of this, I ask that you send me a very brief e-mail before the lecture (at the latest by 6:00pm the day before) with one question or comment about the chapter. This should be no more than a few sentences. This could describe something you found unclear, something you would like to know more about, a comment on something you found particularly bad/good/interesting, etc. I am requiring this both as an incentive for you to look at the chapter in advance and as guidance to me about what to emphasize in the lecture. In light of the surprisingly high enrolment, please help me out by making the subject line of your e-mail STAT 530 Ch. X: Y, where X is the chapter in question and Y is your name.

With everyone having looked at the material in advance, I hope and expect we can have lively in-class discussion!

Assignments: Questions (Assignment 1, Assignment 2) will be posted cumulatively. The first assignment will be due Friday February 12 (the last day before the Olympics-extended reading break) and the second assignment will be due Thursday April 1 (in advance of the 4-day Easter weekend).

Term paper: Details now posted here. You will have a lot of flexibility in choosing to write about some aspect of Bayesian inference of interest to you, and there will be stated limits to ensure this is a short term paper.





Contact Information Link Image