Samuel D. Oman


Department of Statistics
Hebrew University of Jerusalem
Mt. Scopus, Jerusalem, ISRAEL
Phone: +972-2-588-3442 
Fax: +972-2-588-3549
Email: oman@mscc.huji.ac.il




Explaining the Bootstrap

IBEE Computer Code

52542 Generalized Linear Models

     IBEE

   The following two files describe and contain the S-plus code used to compute the IBEE
and Independence estimates for the Meron vegetation data in "Analyzing Spatially
Distributed Binary Data Using Independent-Block Estimating Equations" by Oman,
Landsman, Carmel and Kadmon.
    The code may be used for binary responses from an equally-spaced rectangular grid,
and uses an exponential covariance function for the latent normal field (it may be
easily adapted to handle other types of covariance functions).

    The S-plus statements are in a number of programs ( ibee.1 - ibee.6) which are to be
run sequentially.  Before running them, the user must prepare data matrices and define
a number of parameters, as described in the comments of  ibee.1.  The programs use a
number of functions, whose code is also available.

ibee.readme.rtf    Contains a brief description of the programs, the names of the functions,
                            and the calling sequence for the functions.

ibee.progs.rtf    Contains the code for ibee.1 - ibee.6 and the functions, as one long file. 
                        The first line of each program or function is in boldface, and the last line
                        is followed by a blank line, in order to simplify copying and pasting into
                        the Splus window.  The programs and functions have numerous
                       comments.


Generalized Linear Models

Week 1
syllabus.rtf                     Course syllabus.
glim.formulas.pdf          Some formulas (exponential family, derivatives of log-likelihood, ...).

medfly.doc                     Description of data set on medfly trappings.
medfly.data.txt               Medfly data set.
medfly.data.tab.txt         Medfly data set; tab-delimited and contains variable names. 

meron.1.doc                  Description of data set on vegetation growth in Mt Meron.
meron.200.glim.txt       Mt Meron data set.

Week 2
exercise.0.soln.pdf           Solution to exercise 0.
exercise.1.doc                  Exercise 1.
mvf.pdf                            Derivation of formulas for E(Y) and var(Y) in terms of the natural parameter θ\theta.

Week 3
medfly.explor.R               R code for exploratory analysis of medfly data.
medfly.explor.doc            Results of exploratory analysis.

Week 5
exercise.1.soln.pdf               Solution to Exercise 1.
medfly.poisson.2.R              Code to fit a Poisson regression to the medfly data.
medfly.poisson.rslts.pdf       Results of the fit.
mismatch.data.txt                 Data on survival time following heart transplant.
glim.mismatch.fit.R             R code to analyze the data.
mismatch.gamma.1.doc       Results of the analysis.

Week 7
exercise.2.rtf              Exercise 2.
digamma.pdf              Some graphs with the digamma function.

Week 9
exercise.2.soln              Solution to Exercise 2.
meron.analysis.1.R       Code for the solution.
wald.score.lr.doc           Graph showing the relationship between the LR, Wald and score tests.
exercise.3.rtf                  Exercise 3.

Week 10
exercise.3.soln.R         Code to solve Exercise 3.
exercise.3.soln.pdf      And the results.

Week 11
dist.examples.pdf        Examples of distributions with their link functions.  From McCullagh and Nelder
                                         (see syllabus).
mc.int.exercise.pdf      An exercise illustrating the difference between multicollinearity and interaction,
                                           in the context of linear regression.
mc.int.soln.pdf            Solution to the exercise.
exam.guidelines.pdf    Guidelines for the exam.
past.exam.pdf              A past exam.
exam.solution.pdf       And its answers.
aic.examples.pdf         Examples of model selection using AIC.

Week 12
mismatch.lm.diagnostics.R               Code to compute linear-model diagnostics for the mismatch data.
mismatch.glm.diagnostics.R             Code to compute GLIM diagnostics for the mismatch data.
mismatch.diagnostics.rslts.1.pdf       Diagnostics for the mismatch data.
mismatch.diagnostics.rslts.2.pdf       More diagnostics.

Week 13
overdispersion.examples.pdf       Examples of overdispersion.
medfly.quasi.R                             Code to analyze the medfly data using quasi-likelihood.
medfly.quasi.rslts.pdf                   And the results.

For the exam
more.formulas.pdf                 An additional page of formulas which will be attached to the exam (together with those from Week 1).
moed.a.solution.pdf               Solution to moed aleph.
moed.a.stem.leaf.pdf             Stem-and-leaf diagram of grades for moed aleph.