| gfcure {gfcure} | R Documentation |
Fit a parametric accelerated failure time mixture cure model with distributions from the generalized F distribution family. It can also be used to fit the AFT model without the cure portion.
gfcure(formula = formula(data), cureform = NULL, data = sys.parent(),
subset, na.action, init = NULL, dist = c("weibull", "lognormal", "gamma",
"gll", "loglogistic", "gf", "egg", "exponential", "rayleigh"), sait = 100,
ntit = 20, eps = 1e-04, temp = 0, ntemp = 0, fixed = NULL,
debug = c("NR", "NM", "GFCURE"), evl = F, model = F, x = F, z = F,
y = T)
formula |
a formula expression similar to the one used in survreg. The response is a survival object returned by the Surv() function. It is used to specify the covariate effects on the failure time of uncured subjects. See the documentation for survreg, Surv for details. The expression to the right of the "~" specifies the effect of covariates on the failure time of uncured patients. |
cureform |
a formula expression similar to formula except that it does not contain a response. It is used to specify the effects of covariates on the cure rate. A covariate may be used in both formula and cureform. A model with the intercept term only assumes that there are cured patients and that the cure rates are the same for all patients. If cureform is not given, gfcure fits an AFT model to data |
data |
a data frame in which to interpret the variables named in the formula, cureform, or the subset argument. |
subset |
expression saying that only a subset of the rows of the data should be used in the fit. |
na.action |
a missing-data filter function, applied to the model.frame, after any subset argument has been used. Default is options()$na.action. |
init |
a vector of initial values of the parameters in the parametric mixture model. It is optional. |
dist |
assumed distribution for the survival time. It can be one of "weibull", "lognormal", "gamma", "gll", "loglogistic", "gf", "egg", "exponential" and "rayleigh", where "gll", "gf" and "egg" respectively stand for generalized log-logistic, generalized F and extended generalized gamma distributions. The default is "weibull". |
sait, ntit |
the number of iterations in the simulated annealing algorithm (or the Nelder-Mead simplex algorithm) and the Newton-Raphson algorithm. Note that the simulated annealing algorithm is only used for a distribution with more than two parameters, such as the generalized F and the EGG distributions. |
eps |
tolerance for convergence. Iteration stops once the relative change in deviance is less than eps. |
temp, ntemp |
a maximum temperature and a number of times the temperature is lowered from the maximum to zero in the simulated annealing algorithm. If either of them is zero, the simulated annealing algorithm reduces to the Nelder-Mead simplex algorithm. |
evl |
a logical value, if TRUE, then the log-likelihood function is evaluated at given values of parameters in init. |
model |
a logical value, if TRUE, then the model frame is returned only. |
x |
If TRUE, the covariate matrix in the accelerated failure time part of the parametric mixture model is added to the final result. |
z |
If TRUE, the covariate matrix in the logistic part of the parametric mixture model is added to the final result. |
y |
If TRUE, the Surv object used in the fit is added to the final result. |
A mixture cure model is considered in this function. For cure rate, a logistic regression model is employed and the probability of being cured is given by (1+exp(gamma Z))^(-1). For uncured subject, the failure time is modeled by an accelerated failure time model with distributions from the generalized F distribution family. The covariate effects on time is modeled through the scale parameter of the distribution.
A covariate can be used either in formula or in cureform or in both.
The model parameters are estimated by maximizing the observed log-likelihood function directly and the standard error estimates are obtained from minues the second derivatives of the log-likelihood function.
We note that the maximum number of parameters estimable in this program is 174, which is set in MAXCOV in gfcure.h. Increasing this value may cause the failure of this program.
an object of class gfcure is returned. It can be examined by print(), summary() and predict().
call |
The call to gfcure |
coef |
The estimated parameters in the model |
loglik |
The maximum log-likelihood |
infomat |
The information matrix from the log-likelihood |
ncoefx |
The number of covariate parameters estimated from the AFT model part for uncured subjects, including intercept |
ncoefz |
The number of parameters estimated from the logistic regression model part for cure rate, including intercept |
nshape |
The number of shape parameters in the distribution for the AFT model |
nscale |
The number of scale parameters in the distribution for the AFT model |
iter |
The number of iterations in maximizing the observed log-likelihood function |
dist |
The name of the distribution used in the model |
Yingwei Peng
Peng, Y., Dear, K.B.G. and Denham, J. W. "A generalized F mixture model for cure rate estimation," Statistics in Medicine 17, 813–830 (1998).
data(simdata) gfcure(Surv(time, cens) ~ x1, ~ x1, data = simdata)