Displaying 1 result from an estimated 1 matches for "cox10".
Did you mean:
col10
2018 Mar 28
0
coxme in R underestimates variance of random effect, when random effect is on observation level
...one")
### Same as previous example, but patients are grouped
sd.10<-rep(0,50)
for (i in 1:50){
data10<-simulWeib.group(25000,lambda=0.0001,rho=2,beta1=0.33,beta2=5,beta3=0.25,beta4=0,rateC=0.0000000001, sigma = 0.25, M=40)
data10$group<-as.factor(data10$group)
fit.cox10<-coxme(Surv(time,status) ~ x1 + x2 + x3 + (1 | group), data=data10)
sd.10[i]<-sqrt(as.numeric(fit.cox10$vcoef))
print(i)
}
print("model 10 done")
PhD student
Health e-Research Centre ? Farr Institute
Rm 2.006 ? Vaughan House ? Portsmouth Street ? University of Manche...