New! The second homework is due Monday, Oct 10.
date | R/Sweave source | R/Sweave output | description |
---|---|---|---|
Sep 07 | gofmc.R | gofmc.Rout | Good Old-Fashioned Monte Carlo |
Sep 09 | demo.Rnw | demo.pdf | R mcmc package vignette.
|
Sep 09 | demo-new.Rnw | demo-new.pdf |
Revised R mcmc package vignette.
|
Sep 19 | (none) | theo.pdf | Theory notes (unfinished). Last changed Fri Oct 28 16:10:32 CDT 2005. |
Nov 02 | (none) | bmhg.pdf | New! Notes about the Metropolis-Hastings-Green algorithm. |
Nov 16 | (none) | cleo.pdf | New! More theory notes (unfinished). Last changed Wed Nov 16 13:17 CST 2005. |
Nov 21 | (none) | n1995.pdf | New! More MCMC notes (unfinished, from 1995). Last changed Wed Nov 21 10:50 CST 2005. |
Nov 21 | (none) | n1998.pdf | New! More MCMC notes (unfinished, from 1998). Last changed Wed Nov 21 11:15 CST 2005. |
Handed out in class Sep 16.
Handed out in class Sep 26. Due Monday, Oct 10.
Handed out in class Oct 12. Due Friday Oct 21.
The source file for this spin.Rnw will save retyping, even if you do not use Sweave. The data file for this betas.txt will guarantee the same results (modulo Monte Carlo error even if R's random number generators should change in a new version).
Handed out in class Oct 24, Due Wednesday Nov 8.
Revised Oct 25 (summary of model added in Section 1.1 and required MCSE changed from 0.001 to 0.25).
The source file for this flu.Rnw will save retyping, even if you do not use Sweave. The data file for this is flu.txt.
Handed out in class Dec 5. Last homework. Due by the end of the semester. This is a theory homework, no computing. There is a brief hint in the homework assignment.
A more extended hint giving a specific drift function to use is in Homework 5 Hint, but you should not look at that if you don't need it.
My solution to homework 1 as an R batch file hw1a.R and as the results of running
R CMD BATCH --vanilla hw1a.R
on it hw1a.Rout.
My solution to homework 2 as an R batch file hw2s.Rnw which produces the PDF hw2s.pdf.
My solutions to homework 3 as R batch files spin1.Rnw and spin2.Rnw which produce the PDFs spin1.pdf and spin2.pdf.
I never went back and corrected the mistake in the sampler.
My solution to homework 4 as
an R batch file flu1.Rnw
which produces the PDF flu1.pdf.
(although the file says Part I
, there is no part 2).
A paper Markov chain Monte Carlo maximum likelihood from the 1991 Interface Proceedings that is hard to find and has the first mention of what is now called parallel tempering as PostScript (the original form) or as PDF (which may not look very good).