An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20090802/0a90d3b0/attachment-0001.pl>
Competing Risks Regression with qualitative predictor with more than 2 categories
2 messages · kende jan, Ravi Varadhan
Hi,
You can use `model.matrix' to create the apropriate design matrix for factor variables.
set.seed(10)
ftime <- rexp(200)
fstatus <- sample(0:2,200,replace=TRUE)
gg <- factor(sample(1:3,200,replace=TRUE),1:3, c('a','b','c'))
cov <- matrix(runif(600),nrow=200)
dimnames(cov)[[2]] <- c('x1','x2','x3')
cov2 = model.matrix( ~ cov + gg)
print(z <- crr(ftime,fstatus,cov2[, -1])) # you shouldn't have intercept in the FG model
Hope this helps,
Ravi.
____________________________________________________________________
Ravi Varadhan, Ph.D.
Assistant Professor,
Division of Geriatric Medicine and Gerontology
School of Medicine
Johns Hopkins University
Ph. (410) 502-2619
email: rvaradhan at jhmi.edu
----- Original Message -----
From: kende jan <kendejan at yahoo.fr>
Date: Sunday, August 2, 2009 6:01 am
Subject: [R] Competing Risks Regression with qualitative predictor with more than 2 categories
To: r-help at r-project.org
Hello, I have a question regarding competing risk regression using cmprsk package (function crr()). I am using R2.9.1. How can I do to assess the effect of qualitative predictor (gg) with more than two categories (a,b,c) categorie c is the reference category. See above results, gg
[[elided Yahoo spam]]
Thank you for your help Jan
> # simulated data to test
> set.seed(10)
> ftime <- rexp(200)
> fstatus <- sample(0:2,200,replace=TRUE)
> gg <- factor(sample(1:3,200,replace=TRUE),1:3,c('a','b','c'))
> cov <- matrix(runif(600),nrow=200)
> dimnames(cov)[[2]] <- c('x1','x2','x3')
> cov2=cbind(cov,gg)
> print(z <- crr(ftime,fstatus,cov2))
convergence: TRUE
coefficients:
x1 x2 x3 gg
0.2624 0.6515 -0.8745 -0.1144
standard errors:
[1] 0.3839 0.3964 0.4559 0.1452
two-sided p-values:
x1 x2 x3 gg
0.490 0.100 0.055 0.430
> summary(z)
Competing Risks Regression
Call:
crr(ftime = ftime, fstatus = fstatus, cov1 = cov2)
coef exp(coef) se(coef) z p-value
x1 0.262 1.300 0.384 0.683 0.490
x2 0.652 1.918 0.396 1.643 0.100
x3 -0.874 0.417 0.456 -1.918 0.055
gg -0.114 0.892 0.145 -0.788 0.430
[[alternative HTML version deleted]]
______________________________________________ R-help at r-project.org mailing list PLEASE do read the posting guide and provide commented, minimal, self-contained, reproducible code.