Statistics 694: Applied Bayesian Analysis -- Winter 2007
Instructor
Dr. Catherine Calder
Grader
Mr. Yongku Kim
Course Description
This course aims to provide a practical
introduction to Bayesian data analysis. Students will be exposed to a
variety of Bayesian models including the Bayesian linear model for
normal and non-normal data. Bayesian hierarchical modeling will be
discussed as a strategy for modeling complex processes and as a means
of assimilating a variety of sources of data, and students will be
required to complete a project in which they fit a hierarchical model
to data.Simulation-based methods for model-fitting will be introduced,
and students will learn to use the WinBUGS/OpenBUGS software in
addition to programming basic MCMC algorithms in R.
Prerequisites
Statistical
Theory (Stat 520 and 521, Stat 610 and 623, or Stat 620 and 621) and
Applied Regression Analysis (Stat 645), or permission of the
instructor
Website http://www.stat.ohio-state.edu/~calder/stat694-wi07/
Important announcements, lecture notes, homework problems and
solutions, computing references, and other information about the class
are posted on the course website.
Lectures TTh
10:00-11:18am in 312 Cockins Hall
(CH)
Lecture notes will be posted on the course website before
class. Please read the sections of the textbook that will be covered,
and print out a copy of the lecture notes before each class. There
may be parts of the notes that you should fill in during lecture, and
you may need to take separate notes on examples that are not in the
lecture notes. Unless instructed otherwise, you are responsible for
all of the material in the sections of the book that are covered in
lecture even if some of the material in the book section is not
covered in class. If you are unsure if you are responsible for a
particular topic, be sure to ask the instructor.
Required Textbook
Gelman, A., Carlin, J.B., Stern, H., and Rubin, D. Bayesian Data Analysis, Second Edition. Chapman and Hall, 2004.
Midterm
Exam
There will be an in-class midterm given on Thursday,
Feb. 8th. Re-grade requests on the midterm exam must be
submitted to the grader in writing within one week of the day the
midterms are handed back. Please bring a calculator to the
midterm exam.
Project
Each
student is required to independently complete a project involving
fitting a hierarchical Bayesian model to real data. A detailed
description of the requirements for the project proposal (due on
Thursday, Feb. 22nd) and report (due on Tuesday, March
13th by 5pm) will be distributed in class and on the course
website.
Homework
Assignments
There will be four homework assignments for the
course. You are encouraged to work together on the problems, but each
student must hand in his or her own work. DO NOT COPY any part
of another student's homework including computer output.
Solutions to the homework problems will be posted on the course
website. Late homework assignments will be accepted until the
solutions have been posted on the website. Once the solutions have
been posted, late homework will not be accepted. If you are unable to
come to class the day a homework assignment is due, please contact the
instructor. Re-grade requests on the homework problems must be
submitted in writing to the course grader within one week of the day
the solutions are posted.
Grading
The following is a breakdown of your final course grade:
| Midterm | 35% |
| Project | 35% |
| Homework | 30% |
Grades on the midterm exam may be curved if necessary.
Computing
We
will be using the R statistical computing package and the Bayesian
Inference Using Gibbs Sampling (WinBUGS/OpenBUGS) software, which are
both freely available. No prior knowledge of these computing packages
is required, although experience with R (or S-plus) will be helpful.
Both R and WinBUGS/OpenBUGS are available in the Department of
Statistics computing laboratory, although this facility is only
available to Statistics students. Links to websites where these tools
can be downloaded and reference manuals are available on the course
website. Most homework assignments will require some computing.
Please cut and paste your computer output and graphs into your
homework solutions.
Special
Accommodations
If you need any accommodations based on the
impact of a documented disability contact the instructor privately to
discuss your specific needs. You should also contact the Office of
Disability Services to coordinate special accommodations.
Academic
Misconduct
Academic misconduct will not be tolerated and
will be dealt with procedurally in accordance with university
policy.
