| censor.x {coxed} | R Documentation |
Generate right-censoring to be dependent on the data
Description
This function is called by sim.survdata and is not intended to be used by itself.
Usage
censor.x(x, censor = 0.1)
Arguments
x |
A matrix or data frame containing covariates |
censor |
The proportion of observations to designate as being right-censored |
Details
The purpose of this function is to efficiently generate indicators for whether or not an observation in simulated duration data is right-censored. In this case, whether or not an observation is right-censored depends on the covariates in the data.
This function randomly draws new coefficients, one for every column of x, from the normal distribution with mean 0 and
standard deviation of 0.1. It uses these new coefficients to build a linear predictor, to which is added a disturbance term which
is also drawn from N(0,.1). A fixed proportion, given by censor, of the observations with the highest values of this linear
predictor are set to be TRUE and the others are set to FALSE.
Value
A vector of logical values
Author(s)
Jonathan Kropko <jkropko@virginia.edu> and Jeffrey J. Harden <jharden2@nd.edu>
See Also
Examples
Xdata <- matrix(rnorm(300), 100, 3)
censor.x(Xdata, .1)