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')