simSurv: The function that simulates independent/cluster-correlated...

In SemiCompRisks: Hierarchical Models for Parametric and Semi-Parametric Analyses of Semi-Competing Risks Data

Description

The function to simulate independent/cluster-correlated right-censored survival data under Weibull/Weibull-Normal model.

Usage

simSurv(id=NULL, x, beta.true, alpha.true, kappa.true, sigmaV.true=NULL, cens)

Arguments

id	a vector of cluster information for n subjects. The cluster membership must be set to consecutive positive integers, 1:J. Required only when generating clustered data.
x	covariate matrix, n observations by p variables.
beta.true	true value for β.
alpha.true	true value for α.
kappa.true	true value for κ.
sigmaV.true	true value for σ_V. Required only when generating clustered data.
cens	a vector with two numeric elements. The right censoring times are generated from Uniform(cens[1], cens[2]).

Value

simSurv returns a data.frame containing univariate time-to-event outcomes from n subjects. It is of dimension n\times 2: the columns correspond to y, δ.

y	a vector of n times to the event
delta	a vector of n censoring indicators for the event time (1=event occurred, 0=censored)

Author(s)

Kyu Ha Lee and Sebastien Haneuse
Maintainer: Kyu Ha Lee <[e-mail]>

Examples

	set.seed(123456)
	
	J = 110
	nj = 50
	n = J * nj

	id <- rep(1:J, each = nj)

	x	= matrix(0, n, 2)	
	x[,1]	= rnorm(n, 0, 2)	
	x[,2]	= sample(c(0, 1), n, replace = TRUE)

	beta.true = c(0.5, 0.5)
	
	alpha.true = 1.5		
	kappa.true = 0.02
	sigmaV.true = 0.1

	cens <- c(30, 40)		

	simData <- simSurv(id, x, beta.true, alpha.true, kappa.true, 
				sigmaV.true, cens) 		 

SemiCompRisks documentation built on Feb. 3, 2021, 5:06 p.m.