****************************************************************
One is choosing a textbook for the class. I have skimmed the
Kennedy and Gentle text (a 1980 edition). I have heard of the text by
Thisted but unfortunately do not yet have access to it (but will read
the several reviews, eg, JASA '89, that are available). I have found
Devroye's Non-Uniform Random Variate text and Ripley's Stochastic
Simulation text very valueable in my own work.
Question #1- what written materials have people used for such a course,
and what do they like/dislike/recommend?
****************************************************************
A second topic is course coverage. This should perhaps go before
talking about a book! I'm an admitted novice at teaching this and
am somewhat self-taught. My thoughts so far have been to at least discuss
1. random variate generation- techniques for r. uniform generators,
evaluating r. uniform generators, techniques for non-uniform rv's
2. computer intensive estimation- bootstrap CIs, Markov Chain Monte Carlo
(esp Hastings' algorithm), EM algo, iterated bootstrap
3. computer intensive testing- Monte Carlo tests, bootstrap CIs,
randomization tests
4. some on numerical methods and computers- optimization approaches
and issues (eg, scaling), Monte Carlo integration
This is undoubtedly quite incomplete and there is overlap in the categories
(eg, I've left out smoothers, iteratively reweighted least squares, much
of the graphical work)- I'm finding it hard to some degree to see what is not
statistical computing anymore!
Question #2- what do people like as a course outline, what is emphasized?
****************************************************************
And a third topic is the hands-on experience for the students. My
own experience has been to use S-Plus combined with C pgms, often
borrowed from Numerical Recipes, and occasionally call in canned FTN
routines.
Question #3- what kinds of hands-on work do you like your students to do,
and how?
If you would, please send your comments via email to me. Thanks.
-- Ken Newman (ken@laplace.mathstat.uidaho) Dept of Mathematics and Statistics Voice: (208)-885-6861 University of Idaho FAX : (208)-885-5843 Moscow, ID Telepathy: Think about it