Skip to contents

Random number generation for Correlated Survival and Ordinal Outcomes

generate_rv.Rd

This function generates random number vectors of correlated survival and ordinal outcomes. Gaussian or Clayton copula is used to specify the correlation structure. Marginal distributions are based on a exponential distribution. The order outcome is generated by discretizing one of the correlated random numbers based on the quantile.

Usage

generate_rv(
  target_tau,
  num_samples,
  censoring_rate,
  copula_type = c("Gaussian", "Clayton"),
  x_levels = 5,
  x_probs = c("equal", "random"),
  x_values = NULL,
  hr = 1
)

Arguments

target_tau

the target value of the Kendall's \(\tau\)

num_samples

the number of simulation samples

censoring_rate

the censoring rate

copula_type

copula type (default = "Gaussian")

  • Gaussian: Gaussian copula

  • Clayton: Clayton copula

x_levels

the number of category for the ordinal outcome (default = 5)

x_probs

the proportion of each level of x (default = "equal")

  • equal: equal sample sizes

  • random: random sample sizes

x_values

the cutoff values of x. If x_values is specified, x_probs is ignored. x_levels - 1 and the lengths of x_levels must be equal.

hr

the hazard rate for survival time (default = 1.0)

Value

  • y: the survival time or censoring time outcome vector

  • event: the event indicator outcome vector

  • x: the ordinal (or continuous) outcome vector

  • t: the true survival time vector (for simulation)

  • c: the true censoring time vector (for simulation)

  • x0: x before discretization (for simulation)

Examples

library(surrosurvo)
set.seed(1234)
data <- generate_rv(0.7, 500, 0.3)
surrosurvo(data$y, data$event, data$x)
#>   method       tau         se       lcl       ucl
#> 1 taumo1 0.6556501 0.02496919 0.6067114 0.7045888
#> 2 taumo2 0.6563674 0.01726476 0.6225291 0.6902057
#> 3   tauo 0.5042325 0.01834774 0.4682716 0.5401934
#> 4  tauso 0.7317075 0.02600037 0.6807477 0.7826672

# 5-level ordinal variable with Clayton copula
data <- generate_rv(0.7, 500, 0.1, copula_type = "Clayton", x_levels = 5)
surrosurvo(data$y, data$event, data$x)
#>   method       tau         se       lcl       ucl
#> 1 taumo1 0.6375091 0.01774413 0.6027312 0.6722869
#> 2 taumo2 0.6376534 0.01599582 0.6063021 0.6690046
#> 3   tauo 0.5999920 0.01633828 0.5679696 0.6320144
#> 4  tauso 0.7182260 0.01822755 0.6825006 0.7539513