The PTCMGH R package implements promotion time cure models of the type:
If a vector of covariates
-
Let
${\bf w} \subseteq {\bf x}$ , and consider the log-link$\theta({\bf w}) = \exp({\bf w}^{\top}{\boldsymbol{\alpha}})$ . -
Let
${\bf z} \subseteq {\bf x}$ , and consider the hazard structure$\tilde{H}(t \mid {\bf z})$ . Then, we can define
The PTCMGH R package provides functions for simulating survival times from PTCMs with proportional hazards, accelerated failure time, and general hazard structures under a log-link for
-
General Hazard (GH) model.
-
Accelerated Failure Time (AFT) model.
-
Proportional Hazards (PH) model.
-
Accelerated Hazards (AH) model.
These models are fitted using the R commands nlminb and optim. Thus, the user needs to specify the initial points and to check the convergence of the optimisation step, as usual.
The current version of the PTCMGH R package implements the following parametric baseline hazards for the models discussed in the previous section, using the command PTCMMLE.
-
Power Generalised Weibull (PGW) distribution.
-
Exponentiated Weibull (EW) distribution.
-
Generalised Gamma (GenGamma) distribuiton.
-
Gamma (Gamma) distribution.
-
Lognormal (LogNormal) distribution.
-
Log-logistic (LogLogistic) distribution.
-
Weibull (Weibull) distribution. (only for AFT, PH, and AH models)
All positive parameters are transformed into the real line using a log link (reparameterisation).
The package depends on functions from the HazReg R package, which must also be installed and loaded.
library(devtools)
install_github("FJRubio67/HazReg")
install_github("FJRubio67/PTCMGH")
library(HazReg)
library(PTCMGH)
?simPTCMGH
?PTCMMLE