Download Assignment 4. It is due by Nov 25, 2005.
Download Assignment 3. It is due by Nov 08, 2005.

• Teacher: Arnab Chakraborty
• Class room: 705
• Lab:(for non-B.Stat.s only) 5-th floor computer lab in New Academic Building
• Class hours: Tue, 11:10--12:55, Thur, 3:10--4:55
• Lab hour:(for non-B.Stat.s only) Mon, 5--6pm.

1. Statistical inference using simulation
2. Bootstrapping
3. Cross-validation
4. Jackknifing
5. Markov Chain Monte Carlo techniques

The main reference material is the set of online classnotes. These contain both the class lectures (more or less) as well as programming tips. You may also like to read the book "Introduction to the bootstrap" by Efron and Tibshirani for the topics bootstrapping, cross-validation and jackknife. The recent book "Statistical Computing: Existing Results and Recent Trends" by D. Kundu and A. Basu is also a good reference for the same topics.

I do not know of any single good book for Markov Chain Monte Carlo. Here is a relevant list of publications.

We shall use Matlab and SAS in this course. You may use either Matlab 6.5 or 7. For SAS we shall use version 9.1.3.

Local links:

• Getting started with SAS: This is written to warm you up with the SAS environment that we have at ISI.
• Working with SAS: This is a sequel to the above.
• A closer look at SAS: This is the third instalment in the above sequel.
• MCMC with SAS: This is the fourth instalment in the above sequel.

Internet links:

• Online SAS documentation (Version 8): We shall not use version 8, but this documentation is still helpful.
• Online SAS documentation (Version 9.1.3) This is the version we shall use. But it is less friendly than the documentation for version 8.
• Help on SAS/IML: SAS has a command called "PROC IML". Once you issue this command, SAS becomes a powerful programming language (somewhat like Matlab). This page gives some important IML commands.

• A very friendly Matlab primer : This is a link to a local copy that you can read even without connection to the internet.
• A Matlab primer

Click on the links below to read the class notes online. If you want to print the notes then use the printer-friendly versions:

• Printer-friendly notes before the Puja.
• Printer-friendly notes after the Puja.

1. Simulation: [ 1 | 2 | 3 ]
2. Bootstrap: [ 1 | 2 | 3 ]
3. Cross validation: [ 1 ]
4. Jackknife: [ 1 ]
5. Markov chains: [ 1 | 2 ]
6. Markov Chain Monte Carlo (MCMC): [ 1 ]
7. Gibbs sampler: [ 1 |  2 |  3 |  4 |  5 |  6 ]
8. Metropolis-Hastings: [ 1 | 2 ]
9. Convergence issues: [ 1 ]