This may be of interest to many participants in the Applied Statistics
Workshop, so I am forwarding it along. A good summer to all. Best, Dan
______________________________
To Whomever May Be Interested:
This email provides details regarding a 12-week informal course on
computer programming specifically designed for fast execution of iterative
statistical techniques (such as MCMC and the EM algorithm). The informal
course is open to all Harvard students and affiliates, although there are
substantive prerequisites as well as a $50.00 fee.
Specifically, on Tuesdays and Thursdays from 10:30 to noon, Dr. Gopi
Goswami of the Harvard Statistics Department will conduct an informal
class on programming in C, C++, and R specifically geared towards
iterative statistical processes. The goal is to teach students how to
program so as to speed up the implementation of, for instance, a Gibbs
sampler, a Metropolis-Hastings algorithm, or an EM algorithm. A basic
introduction to these algorithms is presumed; the course will not review
how and why such methods work, but rather the programming techniques and
languages needed to run them quickly. The class will also presume a basic
facility with the R programming language. Any student who has taken Stat
220, Stat 214, Gov 2003, or the substantial equivalent has the necessary
background knowledge. The class will include a textbook (Deitel &
Deitel's _C++: How To Program_; see below for additional references)
and programming assignments (i.e., homework). Students will complete some
projects in a group format as well as edit one another's computer code.
No formal university credit will be given for the course, and no grades
will be assigned.
Details regarding timing and location appear below.
Each student participating in the class, even those who intend merely to
sit in on a few sessions, will be asked to pay a $50.00 fee.
We hope all interested in this subject matter will take advantage of the
opportunity this informal class represents. Persons having questions
should contact Jim Greiner at greiner(a)fas.harvard.edu.
Details:
Time: Tuesday and Thursday, 10:30-noon
First Class: June 14
Last Class: September 8
Week Off: Class will not meet the week of August 8 (but will meet July 5)
Meeting location: IQSS (Kirkland Street) Room 22
Fee: $50
Text: Deitel & Deitel's _C++: How To Program_
Prerequisites: Stat 220 or Gov 2003, or substantial equivalent, such as
independent knowledge of R and of Bayesian data analysis techniques
Additional References (students should not purchase): Brian W. Kernighan
& Dennis M. Ritchie, The C Programming Language (2d edition); Bjarne
Stroustrup, The C++ Programming Language (special 3d edition); R-exts.pdf
& R-intro.pdf at
www.r-project.org under the "manuals link"