Metalearners-Benchmark dgps

Simulation data sets for comparing metalearners on estimating CATE in survival outcomes

generate_tutorial_survival_data(
  n,
  p,
  p.b = NULL,
  p.i = NULL,
  f.b = NULL,
  f.i = NULL,
  pi = 0.5,
  beta = 1,
  gamma = 1,
  rho = 2,
  cen.scale = 4,
  cenM = "indX",
  n.mc = 10000,
  t0 = NULL,
  Y.max = NULL,
  dgp = "fcomplex"
)

Arguments

n

The sample size

p

The number of covariates

p.b

The number of variables in the main effect function

p.i

The number of variables in the interaction term

f.b

The function form of the main effects (linear / nonlinear)

f.i

The function form of the treatment-covariate interactions (linear / nonlinear)

pi

The propensity score

beta

The coefficients of variables in the main effect

gamma

The coefficients of variables in the interaction

rho

The shape parameter in Weibull distribution for censoring time

cen.scale

The scale parameter in Weibull distribution for censoring time

cenM

The complexity of censoring mechanism (dependent / independent to covariates)

n.mc

The number of monte carlo draws to estimate the treatment effect with. Default is 10000.

t0

The time of interest

Y.max

The maximum failure time

dgp

The type of DGP

Value

A simulation data set

Examples

# \donttest{
n <- 1000; p <- 25
data <- generate_tutorial_survival_data(n, p,
                                        p.b = 1, p.i = 1,
                                        f.b = "L", f.i = "L",
                                        t0 = 0.2)
cate.true <- generate_tutorial_survival_data(n, p,
                                             p.b = 1, p.i = 1,
                                             f.b = "L", f.i = "L",
                                             t0 = 0.2)$catesp
# }