Statistics 694 Announcements
Note: Please check this page regularly for class
announcements.
3/16 The final course grades are now
available on the registrar's website. If you would like to get your
project back, please send me an email to set up an appointment. Have
a nice break!
2/22 MIDTERM SOLUTIONS: midterm-sol.pdf
2/21 Bin Li's office hour on
Friday, Feb. 24, will be moved to Wednesday, Feb. 22, from 10-11am.
2/16 Bin Li's office hours
tomorrow (Feb. 17th) will be from 1:30-2:30pm, instead of
10:30-11:30am.
2/8 The HW3s are graded. Please
stop by my office if you would like to see how you did before the midterm. If I'm not there, Terry England in Cockins Hall
(408) will have them.
2/7 MIDTERM REVIEW (from
class today)
Questions -- midterm-review.pdf
Solutions -- midterm-review-sol.pdf
2/5 ADDITIONAL OFFICE
HOURS
Monday, 2/6 -- Bin Li, 9:30am-11:30am, 305A Cockins Hall
Monday, 2/6 -- Kate Calder, 1:30-3:30pm, 408A Cockins Hall
2/5 HW 3 ANNOUNCEMENT For exercise 2,
you may either assume that the sigma^2s are known (using an estimate from
the data), or you may use an MCMC algorithm and fit the sigma^2s in
addition to the thetas and mu. For exercise 3, you should assume
that the sigma^2s are unknown.
2/2 If you were not in class
today, please download the description of the final project.
2/1 TYPO in
the R code for the schools example
In the sample.mu function, mu.post.var <-
1/(sum(rep(1,N)/(sigma.sq+tau.sq))) instead of mu.post.var <-
sum(rep(1,N)/(sigma.sq+tau.sq)). Your class notes are correct.
Click here
for a revised version of the code.
1/29 Homework 3 is now available
on the Homework page.
Note: The solutions for HW3 will be posted immediately after class on Tuesday,
Feb. 7th, so that you have time to review them before the midterm
exam. As a result, NO late homeworks will be accepted. In addition,
keep in mind that we will not be able to return the homeworks before
the midterm, so if you would like to compare your answers to the
solutions, please make a copy of them before turning them in.
1/25 The following papers are
optional readings on Markov chain Monte Carlo algorithms. Both are
available from JSTOR.
Casella, G., and George, E.I. (1992). Explaining the
Gibbs Sampler. The American Statistician, 46(3),
167-174.
JSTOR
Link
Chib, S., and Greenberg, E. (1995). Understanding the
Metropolis-Hastings Algorithm. The American Statistician,
49(4), 327-335.
JSTOR
Link
Note: You are NOT responsible for additional material in these
papers beyond what is covered in
lecture or in the textbook reading assignments.
1/17 Solutions to HW1 are now
available on the Homework page.
1/16 The second homework assignment
is posted on the Homework page.
1/8 See the Computing page for
more detailed intructions (thanks to Dale Rhoda!) for installing and configuring Tinn-R.
1/5 The first homework assignment
is posted on the Homework page.