The accelerated failure time model with parametric baseline hazard(s). Allows for stratification with different scale and shape in each stratum, and left truncated and right censored data.
aftreg(
formula = formula(data),
data = parent.frame(),
na.action = getOption("na.action"),
dist = "weibull",
init,
shape = 0,
id,
param = c("lifeAcc", "lifeExp"),
control = list(eps = 1e-08, maxiter = 20, trace = FALSE),
singular.ok = TRUE,
model = FALSE,
x = FALSE,
y = TRUE
)a formula object, with the response on the left of a ~ operator, and the terms on the right. The response must be a survival object as returned by the Surv function.
a data.frame in which to interpret the variables named in the formula.
a missing-data filter function, applied to the model.frame,
after any subset argument has been used. Default is
options()$na.action.
Which distribution? Default is "weibull", with the alternatives
"gompertz", "ev", "loglogistic" and "lognormal". A special case like the
exponential can be obtained by choosing "weibull" in combination with
shape = 1.
vector of initial values of the iteration. Default initial value is zero for all variables.
If positive, the shape parameter is fixed at that value. If
zero or negative, the shape parameter is estimated. Stratification is now
regarded as a meaningful option even if shape is fixed.
If there are more than one spell per individual, it is essential to keep spells together by the id argument. This allows for time-varying covariates.
Which parametrization should be used? The lifeAcc uses
the parametrization given in the vignette, while the lifeExp uses the
same as in the survreg function.
a list with components eps (convergence criterion),
maxiter (maximum number of iterations), and trace (logical,
debug output if TRUE). You can change any component without mention
the other(s).
Not used.
Not used.
Return the design matrix in the model object?
Return the response in the model object?
A list of class c("aftreg", "coxreg") with components
Fitted parameter estimates.
Covariance matrix of the estimates.
Vector of length two; first component is the value at the initial parameter values, the second componet is the maximized value.
The score test statistic (at the initial value).
The estimated linear predictors.
Means of the columns of the design matrix.
Weighted (against exposure time) means of covariates; weighted relative frequencies of levels of factors.
Number of spells in indata (possibly after removal of cases with NA's).
Number of events in data.
Used by extractor functions.
Used by extractor functions.
The Wald test statistic (at the initial value).
The Surv vector.
Logical vector indicating the covariates that are factors.
The covariates.
Total Time at Risk.
List of levels of factors.
The calling formula.
The call.
The method.
Did the optimization converge?
Did the optimization fail? (Is NULL if not).
TRUE if shape was fixed in the estimation.
The parametrization.
The parameterization is different from the one used by
survreg, when param = "lifeAcc". The result
is then true acceleration of time. Then the model is
$$S(t; a, b, \beta, z) = S_0((t / \exp(b - z\beta))^{\exp(a)})$$ S(t; a, b, beta, z) = S_0((t/exp(b - z beta))^exp(a))
where \(S_0\) is some standardized
survivor function. The baseline parameters \(a\) and \(b\) are log shape
and log scale, respectively. This is for the default parametrization.
With the lifeExp parametrization, some signs are changed: $$b - z
beta$$ is changed to $$b + z beta$$. For the Gompertz distribution, the
base parametrization is canonical, a necessity for consistency with
the shape/scale paradigm (this is new in 2.3).
# NOT RUN {
data(mort)
aftreg(Surv(enter, exit, event) ~ ses, param = "lifeExp", data = mort)
# }
Run the code above in your browser using DataCamp Workspace