Hi there,
I am an MS student from Bangladesh. I am doing thesis in my MS degree. In
my research, I am generating data from weibull distribution and my model is
accelerated failure time (AFT) model. I am considering right censoring as
well as covariates. Now I have been facing difficulties to generate
censoring time controlling censoring proportion. I am attaching my codes
here.

Problem. I have generated censoring time using a relationship between scale
and covariates from an article for PH model. But my model is AFT. Is it
authentic to use it here? Please help!!!

Sincerely,
Fabiha
library(survival)
##true coefficients
b0=.25    ## intercept
b1=2
b2=3      ## regression coefficients 
b3=.5
n=30
## covariates
library(MASS)
mu=rep(0,3)
sigma=matrix(.7,nrow=3,ncol=3)+diag(3)*.3
x = mvrnorm(n, mu=mu, Sigma=sigma)
x1=x[,1]
x2=x[,2]
x3=x[,3]
## real survival times
t = rweibull(n, shape=.5, scale=exp((b0+b1*x1+b2*x2+b3*x3)))
## censoring times
c = rweibull(n, shape=.5, scale=exp(-(b0+b1*x1+b2*x2+b3*x3)/.5))
## observed times
time = pmin(t, c)
time 
## censoring indicator
status=as.numeric(c>t)
status
## data frame
data=data.frame(time,status,x1,x2,x3)
data
## model fitting
model=survreg(Surv(time, status)~x1+x2+x3, dist="weibull", data=data)
summary=summary(model)
summary

______________________________________________
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.

Reply via email to