similar to: gls with "crossed heteroscedasticity"

Dear All, I was writing a small wrapper to bootstrap a classification algorithm, but if we generate the indices in the "usual way" as: bootindex <- sample(index, N, replace = TRUE) there is a non-zero probability that all the samples belong to only one class, thus leading to problems in the fitting (or that some classes will end up with only one sample, which will be a problem

Dear All, We are about to purchase the second edition of Agresti's "Categorical Data Analysis" (my old copy of the first ed. of that wonderful book is falling apart). I would appreciate suggestions about other comparable books which, if possible, have examples using R/S code (instead of SAS). Thanks, Ram?n -- Ram?n D?az-Uriarte Bioinformatics Unit Centro Nacional de

Dear All, I think there is a typo in the documentation for coxph (library survival). The help says: eps: convergence threshold. Iteration will continue until the relative change in the log-likelihood is less than eps. Default is .0001. However, if I do "coxph.control()" I get: > coxph.control() $eps [1] 1e-09 So the actual eps being used is not 10-4 but

Dear All, I am using fexact.c on a C++ program I wrote. To minimize dependencies on other files (e.g., to not need to include R.h and ctest.h ---now I only include the R files Boolean.h, Constants.h, and Memory.h), I have re-written all declarations of Sints as ints and, what is potentially more serious, I have re-written the line (lines 329 and 330, in fexact.c on R-1.6.1) /* IMAX is the

Dear All, I have been using the randomForest package for a couple of difficult prediction problems (which also share p >> n). The performance is good, but since all the variables in the data set are used, interpretation of what is going on is not easy, even after looking at variable importance as produced by the randomForest run. I have tried a simple "variable selection"

If an array has missing values in different rows, plotting using the formul= a=20 interface can produce errors. Example: fake.data <- matrix(rep(-100:100, 4), ncol =3D 4) par(mfrow =3D c(1,2)) boxplot(fake.data ~ col(fake.data)) abline(h =3D 0, lty =3D 2) boxplot(as.data.frame(fake.data)) abline(h =3D 0, lty =3D 2) ##### Add the missing data fake.data[190:200, 1] <-

Dear all, I have a piece of code, call it "FA", that will be called thousands of times in a typical run of function "FB". I can: a) define FA as a function outside of FB (in the global environment), and call it; b) define FA as a function inside the body of FB and call it; c) "expand inline" FA inside FB. FA mainly does data frame subsetting, runs svd's, and

Dear All, I have a list, where several components are NULL, and I'd like to obtain that very same list without the NULL components (i.e., I do not want to unlist or otherwise loose the rest of the list structure). I can do that with a loop, but how could I do it without a loop? Thanks, Ram?n -- Ram?n D?az-Uriarte Bioinformatics Unit Centro Nacional de Investigaciones Oncol?gicas (CNIO)

Hi, You should also check my msc.features.select from caMassClass package. It has feature selection algorithm that I found useful in case of mass-spectra data. It performs individual feature selection and/or removes highly correlated neighbor features. Jarek -----Original Message----- From: r-help-bounces at stat.math.ethz.ch [mailto:r-help-bounces at stat.math.ethz.ch] Sent: Friday, January

Dear All, I found Andy Liaw's suggestion about using a NUMA (instead of SMP) kernel when running R on amd64 with > 1 CPU http://finzi.psych.upenn.edu/R/Rhelp02a/archive/35109.html A couple of questions: 1. Is this still the case with the newer dual-core opterons (e.g., the 275 et al., families) running Linux (kernel 2.6)? 2. How does this affect using Rmpi (and snow, papply, et al.)

> x <- matrix(1:10, ncol = 2) > xc <- cor(x[, 2], x[, 0]) > cbind(xc, vector()) Process R floating point exception at Tue Feb 11 19:16:51 2003 > version _ platform i686-pc-linux-gnu arch i686 os linux-gnu system i686, linux-gnu status major 1 minor 6.2 year

Dear All, I'm trying to model heteroscedasticity using a multilevel model. To do so, I make use of the nlme package and the weigths-parameter. Let's say that I hypothesize that the exam score of students (normexam) is influenced by their score on a standardized LR test (standLRT). Students are of course nested in "schools". These variables are contained in the

(Apologies for double posting; I sent this to r-help and was suggested that I ask here which I should probable have done to begin with). *************** Dear All, I've got into the habit of installing R from the precompiled Debian binaries, including many of the packages from the r-cran-* Debian packages, and later building from source (e.g., to link against Goto's BLAS, or to build

Dear all, I am trying to include confidence intervals in a xyplot. This is what I am doing: xyplot(x ~ y|z, alpha = 1,band = "global",panel = panel.locfit) (more specifically, in my case x is a binary response from a logistic regression model) The output plot was fine but it did not include the confidence intervals Anyone knows how to do it? (xYplot did not work either) many thanks

Dear R users, I have changed my R version to the new 1.8.1 and some problems appears when using the previous levelplot code. This is a simple example: a <-1:10 b <-11:20 j <- rnorm(100) grid<-expand.grid(a = a, b = b) levelplot(j~a*b, grid) Normaly in my previous vs this was suffice to produce the levelplot. Now, an empty R graphics device appears with the following error

Dear All, This is not strictly R related (though I would implement the solution in R; besides, being this list so helpful for these kinds of stats questions...). I got a "strange" request from a colleage. He has a bunch (approx. 25000) subjects that belong to one of 12 possible classes. In addition, there are 8 covariates (factors) that can take as values either "absence"

Dear R-help members, I have a question regarding how to use varComb function to specify a variance function for the "weights" in the gls. I need to fit a linear model with heteroscedasticity. The variance function is exp(c0+nu0*W +nu1*W^2) where W is a covariate. Initially I want to use varFunc to define my own variance function following the instruction in the Pinheiro and

Dear R-help members, I have a question regarding how to use varComb function to specify a variance function for the "weights" in the gls. I need to fit a linear model with heteroscedasticity. The variance function is exp(c0+nu0*W +nu1*W^2) where W is a covariate. Initially I want to use varFunc to define my own variance function following the instruction in the Pinheiro and Bates

Dear All, I've got into the habit of installing R from the precompiled Debian binaries, including many of the packages from the r-cran-* Debian packages, and later building from source (e.g., to link against Goto's BLAS, or to build patched versions, etc). I install the newly built R to the very same place (/usr/lib/R). This allows me to build and update R when I wish, AND provides the

Hi, I would like to fit a model for a factorial design that allows for unequal variances in all groups. If I am not mistaken, this can be done in lm by specifying weights. A function intended to specify weights for unequal variance structures is provided in the nlme library with the varIdent function. Is it apropriate to use these weights with lm? If not, is there another possibility to do