CSI 991
Seminar in Computational Statistics:
Statistical Computing with R
Fall, 2008
Fridays 4:00pm -- 5:00pm, Research Building I, Room 301
CSI 991
Section 004
Contacts:
csutton@gmu.edu
jgentle@gmu.edu
Some links on this page are available only to members of the seminar group.
The seminar will be conducted in the form of a workshop.
Participants are encouraged to put
R
on their own laptops, and to bring them
to the seminars.
The required text is
Statistical Computing with R,
by Rizzo.
The R code given in the book is available at
http://personal.bgsu.edu/~mrizzo/SCR/
September 12
Some background on the bootstrap
September 26
Some R code for Monte Carlo studies of the performance of
the bootstrap and the jackknife for estimating the standard
error of a sample median.
October 3
October 10
Last week,
we discussed the question of quantile estimation. The Harrell-Davis
estimator is implemented in the hd function of Rand Wilcox's functions,
available at
an old website for a course, which Jarek found. It is also
implemented in the hdquantile function in
Harrell's Hmisc package. The hdquantile function also uses a leave-out-one
jackknife to provide an estimate of the standard deviation. We may
find it an interesting exercise to compare the jackknife sd estimate
with the bootstrap estimate, using something like
boot(x, function(x,i) hdquantile(x[i], probs=(1:3)/4), R=400)
similar to what we did with the simple sample median.