Simulate Survival Data — rsurv • ipred

Simulation Setup for Survival Data.

Usage

rsurv(N, model=c("A", "B", "C", "D", "tree"), gamma=NULL, fact=1, pnon=10,
      gethaz=FALSE)

Arguments

N: number of observations.
model: type of model.
gamma: simulate censoring time as runif(N, 0, gamma). Defaults to NULL (no censoring).
fact: scale parameter for model=tree.
pnon: number of additional non-informative variables for the tree model.
gethaz: logical, indicating wheather the hazard rate for each observation should be returned.

Details

Simulation setup similar to configurations used in LeBlanc and Crowley (1992) or Keles and Segal (2002) as well as a tree model used in Hothorn et al. (2004). See Hothorn et al. (2004) for the details.

Value

A data frame with elements time, cens, X1 ... X5. If pnon > 0, additional noninformative covariables are added. If gethaz=TRUE, the hazard attribute returns the hazard rates.

References

M. LeBlanc and J. Crowley (1992), Relative Risk Trees for Censored Survival Data. Biometrics 48, 411–425.

S. Keles and M. R. Segal (2002), Residual-based tree-structured survival analysis. Statistics in Medicine, 21, 313–326.

Torsten Hothorn, Berthold Lausen, Axel Benner and Martin Radespiel-Troeger (2004), Bagging Survival Trees. Statistics in Medicine, 23(1), 77–91.

Examples


library("survival")
# 3*X1 + X2
simdat <- rsurv(500, model="C")
coxph(Surv(time, cens) ~ ., data=simdat)
#> Call:
#> coxph(formula = Surv(time, cens) ~ ., data = simdat)
#> 
#>       coef exp(coef) se(coef)      z        p
#> X1  3.2506   25.8050   0.2017 16.115  < 2e-16
#> X2  1.1106    3.0362   0.1657  6.704 2.03e-11
#> X3  0.2247    1.2520   0.1602  1.403    0.161
#> X4  0.0611    1.0630   0.1630  0.375    0.708
#> X5 -0.1698    0.8438   0.1507 -1.127    0.260
#> 
#> Likelihood ratio test=307.4  on 5 df, p=< 2.2e-16
#> n= 500, number of events= 500