[R] clogit & weights
Merry Christmas everyone: I have the following data(mydat) and would like to fit a conditional logistic regression model considering "weights". id case exposure weights 1 1 1 2 1 0 0 2 2 1 1 2 2 0 0 2 3 1 1 1 3 0 0 1 4 1 0 2 4 0 1 2 The R function"clogit" is for such purposes but it ignores weights. I tried function"mclogit" instead which seems that it considers the weights option:##options(scipen=999)library(mclogit)# create the above data frameid = c(1,1,2,2,3,3,4,4)case = c(1,0,1,0,1,0,1,0)exposure = c(1,0,1,0,1,0,0,1)weights = c(2,2,2,2,1,1,2,2)(mydata = data.frame(id,case,exposure,weights)) fit = mclogit(cbind(case,id) ~ exposure,weights=weights, data=mydata)summary(fit)## The answer,however, doesn't seem to be correct. Could anyone pleaseprovides me with some solution to this? Thanks in advance,Keramat Nourijelyani,PhD [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Cross-over Data with Kenward-Roger correction
Dear all:for the folowing data, a two-period, two treatment (A=1 vs. B=2) cross-over is fitted using the folowing SAS code. data one; input Sbj Seq Per Trt PEF; cards; 1 1 1 1 310 1 1 2 2 270 4 1 1 1 310 4 1 2 2 260 6 1 1 1 370 6 1 2 2 300 7 1 1 1 410 7 1 2 2 390 10 1 1 1 250 10 1 2 2 210 11 1 1 1 380 11 1 2 2 350 14 1 1 1 330 14 1 2 2 365 2 2 1 2 370 2 2 2 1 385 3 2 1 2 310 3 2 2 1 400 5 2 1 2 380 5 2 2 1 410 9 2 1 2 290 9 2 2 1 320 12 2 1 2 260 12 2 2 1 340 13 2 1 2 90 13 2 2 1 220 ; run; proc mixed data=one method=reml; class Sbj Per Trt; model PEF = Per Trt /ddfm=kr; repeated Trt / sub=Sbj type=un r; lsmeans Trt / cl alpha=0.05; estimate 'B vs. A' Trt -1 1 / alpha=0.1 cl; run; (where kr option is for Kenward-Roger method).I need to use R to reproduce the results similar to what the above SAS code generates. I have used several R functions including lme, lmer with no success so far.Any advice will be greatly appreciated,Sincerely, Keramat [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Bivariate skew normal cdf; very slow
Dear all, I am calculating the bivariate skew normal cdf in "sn" package using "pmsn" function. Although it is quite convenient ( thanks to prof. Azzalini) but it seems to be slow. For example, it takes about 1 minute in calculation of 100k of such cdf values. I am thinking to write a c++ code for this although not very familiar with it. Any other idea? Thanks in advance, sincerely, Keramat Nourijelyani, PhD Associate Professorof Biostatistics Tehran University of Medical Sciences http://tums.ac.ir/faculties/nourij [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] MCMC with cumulative link models
Hi: Could anyone please let me know where I can find [R] code for implementing MCMC with cumulative link models (i.e. for analysis of ordinal probit or ordinal logistic models). Thanks in advance, Regards, Keramat Nourijelyani, PhD Associate Professorof Biostatistics Tehran University of Medical Sciences http://tums.ac.ir/faculties/nourij [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] MLE for probit regression. How to avoid p=1 or p=0
Dear all: I am writing the following small function for a probit likelihood. As indicated, in order to avoid p=1 or p=0, I defined some precisions. I feel however, that there might be a better way to do this. Any help is greatly appreciated. ## ##set limits to avoid px=0 or px=1 precision1 <- 0.99 precision0 <- 0.01 logpost <- function(par, data){ px <- pnorm(b0 + b1x) # to avoid px=1 or px=0 px[px > precision1] <- precision1 px[px < precision0] <- precision0 loga <- sum( y*log(px)+(1-y)*log(1-px) ) loga } # Best, Keramat Nourijelyani, PhD Associate Professorof Biostatistics Tehran University of Medical Sciences http://tums.ac.ir/faculties/nourij [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] Bayesian Haplotype analysisin R
Dear members: Could you please tell me the package for Bayesian haplotype estimation in R. Currently PHASE is available for such purpose but I need a similar R package. Best regards, Keramat Nourijelyani, PhD Associate Professorof Biostatistics Tehran University of Medical Sciences http://tums.ac.ir/faculties/nourij [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
[R] problem with the MLE of the skew-normal shape parameter
Hi everyone: my likelihood function involves the shape parameter of the skew-normal in addition to other parameters. I used both optim and nlm function to find the MLS's. However, every time that I use a different initial value for the shape parameter, I get a different estimate for it. No such problem with other parameters when I exclude the skew-normal shape parameter. Any advice in this regards will be greatly appreciated. Best, Keramat Nouri [[alternative HTML version deleted]] __ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.