survprep: Prepare data for regsurv()

In jeroenhoogland/regsurv: Regularized parametric survival modeling

Prepare data for regsurv()

Description

Prepare data for regsurv()

Usage

survprep(
  tte,
  delta,
  X,
  model.scale = c("loghazard", "logHazard"),
  time.scale = c("logtime", "time"),
  spline.type = c("rcs", "ns"),
  ntimebasis,
  time.knots = NULL,
  tv = NULL,
  nitimebasis = NULL,
  itime.knots = NULL,
  qpoints = 9,
  scales = NULL,
  shifts = NULL
)

Arguments

tte	numeric vector containing time-to-event data
delta	numeric vector containing event indicator data (1=event, 0=censored)
X	design matrix containing all necessary covariate data
model.scale	choose between log hazard and log cumulative hazard
time.scale	choose between log time and linear time for the time scale of the model
spline.type	type of spline used to model time dependencies (rcs=restricted cubic splines, ns=natural splines (as implemented in the splines2 package))
ntimebasis	integer number of basis columns for the spline representation of the log baseline hazard or log cumulative baseline hazard.
time.knots	numeric vector containing baseline spline knot locations. Defaults to equally spaced knots (if NULL).
tv	numeric vector of column indices for the columns in X with time-varying (i.e. non-proportional) effects.
nitimebasis	number of basis columns for the spline representation of non-proportional covariate effects.
itime.knots	numeric vector containing spline knot locations for time-varying covariate effects. Defaults to equally spaced knots (if NULL).
qpoints	number of quadrature points for numerical integration over time (only applicable for models on the log hazard scale)
scales	only used for internal functions
shifts	only used for internal functions

Value

an object of class survprep

Examples

prep <- survprep(tte=simdata$eventtime,
                 delta=simdata$status,
                 X=as.matrix(simdata[ ,grep("x", names(simdata))]),
                 model.scale="loghazard",
                 time.scale="time",
                 spline.type="rcs",
                 ntimebasis=4,
                 qpoints=9)

jeroenhoogland/regsurv documentation built on March 20, 2023, 3:37 a.m.