Elodie Gillain
2013-Apr-11 15:11 UTC
[R] plotting the cumulative hazard function of a process whose intensity function is specified
I would like to simulate a stochastic process by specifying an intensity
function, say
lambda(t)=0.05 + 3 exp(鈭抰)
I have the following code, but it has a number of problems. One issue I have is
I do not know where the time variable is. I would appreciate your help with this
code.
library("survival")
n <- 5000
lamc<-0.08
k1<-0.05
k2<-3
ld<-function(t){k1+k2}
l<-function(t){k1+k2*exp(-t)}
gendata<-function(n,lamc,k1,k2)
{
re<-rexp(5*n)/(k1+k2)
ru<-runif(5*n)
y<-rep(0,n)
k<-1
for (i in 1:n){
u<-re[k]
}
while (ru[k] <l(u)/ld(u)){
k<-k+1
u<-u+re[k]
y[i]<-u
k<-k+1
}
cen<-rexp(n,rate=lamc)
}
obj<-Surv(floor(pmin(y,cen)),ifelse(y<-cen,1,0))
u<-gendata(n,lamc,k1,k2)
lsurv2<-survfit(Surv(floor(pmin(y,cen)) ~ x, data = u)
plot(lsurv2,fun='cumhaz',xlab='temps',ylab='has',noconf)
[[alternative HTML version deleted]]
