Hi Steve, I think that most people using AIC and related information criteria should read the book: Burnham, K.P., Anderson, D.R., 1998. Model selection and inference: a practical information-theoret
Dear all, A few days ago, a question was asked to S-news on how to compute predicted values given an AR(1) correlation structure for residuals (see below). How can I compute predicted values when I h
Dear all, I would like to estimate odds ratios from a 1-M matched case-control design (1 case and M controls). I know how to do when M=1, and I know that I should use coxph when M > 1, but I don't kn
Dear all, Last week, I posted the following request: Thanks to Brad Biggerstaff, Murray Finkelstein and Terry Therneau for their quick, helpful and efficient solutions, which are basically the same:
Dear S+ users, Has anyone implemented Diggle's test for completely random dropouts (missing values in longitudinal data), as described in: Diggle, P.J., K.-Y. Liang and S.L. Zeger (1994). Analysis of
Dear all, I want to plot a large dataset (25,000 rows). A cluster of points appears in the margin, under the x axis. The same phenomenon occurs with plot() and xyplot(). However, when I make partial
Dear all, I have a list which is the result of by() on a dataframe. Each element of the list is a data frame with the same number of columns (13) and a variable number of lines (1 to 7). The length o
Dear all, Last week, I asked the following question: I got 2 replies from Bert Gunter and Julian Taylor, using similar methods: Bert Gunter: for(i in 1:length(Off2)) Off2[[i]] _ t(Off2[[i]]) matrix(u
Dear S+ users, I would like to add a line in splom trellis graphics to improve the visual perception of trends. It works (nearly) fine when I add a loess curve: splom(~ ozone.data, panel=function(x,
My original posting: ** I would like to add a line in splom trellis graphics to improve the visual perception of trends. It works (nearly) fine when I add a loess curve: splom(~ ozone.data, panel=fun
I just met the same problem and I made a private reply to Willem. It looks like some (many ?) referees don't like the style of Trellis graphs (strip labels, tick marks, grid...). Is is another gap be
Correction: Eric Lecoutre <lecoutre@stat.ucl.ac.be> wrote the software http://www.stat.ucl.ac.be/ISpersonnel/lecoutre/Splus/index.html Best wishes, Renaud -- Dr Renaud Lancelot, vétérinaire CIRAD, D
Dear S+ users, Within a for() loop, I fit a list of models (lme() function), some of them failing to converge or producing errors (singularities, etc.), resulting in an aborted loop. How can I ignore
My original posting was: Many thanks to Bill Dunlap, Brian Ripley, Douglas Bates, Henrik Aalborg Nielsen and Bert Gunter for their helpful replies. Three kinds of solution were suggested: 1. try() 2.
This question is related to lme(). I have a complex fixed-effects structure in a linear mixed effects model (5 covariates and all possible interactions). I want to select the best model according to
Some of you may have noticed that all.combos is a vector of logical values, which is not the expected result. Charles Berry would like to correct: combos is now the expected matrix of logical values.
Dear S+ users, I am using S+ 4.5r2 / win98. I would like to print some meaningful text as a label for a shingle (for example, "Light", "Medium" or "Heavy"), in place of its name and the coloured bar
Hi Edward, Thanks for your reply. I works with factors but it does not do what I want here. My problem is with shingles: dat <- data.frame(x=runif(100), y=runif(100), w=equal.count(rnorm(100), number
Dear S+ users, I want to plot a panel of dotplots to describe 8 mortality rates in 3 animal categories. The patterns are rather different for the 3 animal categories, so I want the diseases to be ord
Dear S+ users, I am working with S+ 4.5 r2 / win98, PC with 64 Mb RAM. I just met a strange problem. All the graphs I plot are on a dark-red background. The problem happened suddenly, both for classi