survReg(formula, data=sys.parent(), weights, subset, na.action,
dist="weibull", init, scale=0, control,
model=F, x=F, y=T, ...)
lm and
formula for details.
The response must be the output from the
Surv function.
subset argument.
na.action=na.exclude
to fit the model without NAs, but pad the returned components of residuals
and fitted values to the length of the original data.
"weibull",
"exponential",
"gaussian",
"logistic",
"lognormal"
and
"loglogistic".
If the argument is a character string, then it is assumed to name an
element from
survReg.distributions (enter
names(survReg.distributions)
for a full list of choices).
Otherwise, it is assumed to be a user defined list conforming to this
standard.
survReg.control.
TRUE, the model frame is returned.
TRUE, then the X matrix is returned.
TRUE, then the y vector (or survival times) is returned.
survReg.control.
"survReg" is returned.
The
survReg function is the successor to the
survreg function from
survival4.
The routine uses a Newton-Raphson iteration with step halving, with provision for general penalized term. Fisher scoring is used for intermediate steps where the information matrix is not positive definite.
survReg(Surv(futime, fustat) ~ ecog.ps + rx, data=ovarian, dist='lognormal')