rsurv is a lightweight package with minimal dependencies for randomly
generating survival times efficiently and flexibly. Survival times can
be randomly drawn from any survival distribution and all code is
vectorized so that each individual can have their own survival function.
Monte Carlo simulation studies that aim to simulate realistic data
generating processes based on empirical data are an intended use case.
The package can be installed from GitHub with:
devtools::install_github("hesim-dev/rsurv")Survival times can be randomly drawn from arbitrary survival curves
using the rsurv() function. Survival curves for multiple individuals
are stored in objects of class survival_curves, which are just data
frames in the format shown here:
library("rsurv")
sc <- example_survival_curves(n = 1000)
head(sc)## time id survival
## 1 0.00000000 1 1.0000000
## 2 0.08333333 1 0.9791166
## 3 0.16666667 1 0.9586694
## 4 0.25000000 1 0.9386491
## 5 0.33333333 1 0.9190469
## 6 0.41666667 1 0.8998541
Random survival times for each individual can then be easily drawn:
sim <- rsurv(sc)
length(sim)## [1] 1000
summary(sim)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.08333 0.83333 1.91667 3.08467 4.16667 24.08333