generate_dat: Generates simulated data

In survival-lumc/CauseSpecCovarMI: Multiple imputation for cause-specific Cox models

Generates simulated data

Description

Generates simulated data

Usage

generate_dat(
  n,
  X_type,
  r,
  ev1_pars,
  ev2_pars,
  rate_cens,
  mech = NULL,
  eta1 = NULL,
  p = NULL,
  mod_type = "latent"
)

Arguments

n	Sample size.
X_type	Either "binary", or "contin"
r	Desired correlation between X and Z
ev1_pars	Named list of ("a1", "h1_0", "b1", "gamm1")
ev2_pars	Named list of ("a2", "h2_0", "b2", "gamm2")
rate_cens	Rate of exponential distributiion for censoring. If 0 means no censoring is applied.
mech	Missingness mechanism, one of: "MAR_GEN", "MAR", "MNAR", "MCAR"
eta1	Only necessary for mech != "MCAR" - degree/direction of assocation between variable responsible for missingness and probability of missingness
p	Proportion of missing values in X
mod_type	Either "latent", weibull times generated from separate weibull distribution, or "total", times generated from sum of cause-specific hazards (using inverse transform method). For educational purposes - both methods yield virtually same results.

Value

Data-frame with missings induced

survival-lumc/CauseSpecCovarMI documentation built on June 16, 2022, 9:51 a.m.