penPHcure.object

a list with elements containing the results of the selected model based on the Akaike information criterion (AIC). See Details.

a list with elements containing the results of the selected model based on the Bayesian Information Criterion (BIC). See Details.

a character string indicating the type of penalty used, either <code>"SCAD"</code> or <code>"LASSO"</code>.

pen.type

a data.frame containing the values of the AIC and BIC criteria for each combination of the tuning parameters.

tuneGrid

a list with elements named <code>CURE</code> and <code>SURV</code>, containing the penalty weights. For more details, see <code><a rd-options="" href="/link/penPHcure?package=penPHcure&version=1.0.2" data-mini-rdoc="penPHcure::penPHcure">penPHcure</a></code>.

pen.weights

the sample size (number of individuals).

the number of unique failure times.

logical value: <code>TRUE</code> in case of tied event times.

isTies

the proportion of censored individuals.

censoring

character string indicating the method used to transform the covariates included in the incidence (cure) component from time-varying to time-invariant. See <code><a rd-options="" href="/link/penPHcure?package=penPHcure&version=1.0.2" data-mini-rdoc="penPHcure::penPHcure">penPHcure</a></code> for more details.

which.X

a formula object with all variables involved in the latency (survival) component of the model.

survform

a formula object with all variables involved in the incidence (survival) component of the model.

cureform

object of class <code>call</code>, with all the specified arguments.

call

This class of objects is returned by the function <code><a rd-options="" href="/link/penPHcure?package=penPHcure&version=1.0.2" data-mini-rdoc="penPHcure::penPHcure">penPHcure</a></code> when is called with the argument <code>pen.type = "SCAD" | "LASSO"</code>. Objects of this class have methods for the functions <code>summary</code> and <code>predict</code>.

Implementation of the semi-parametric proportional-hazards (PH) of Sy and Taylor (2000) <doi:10.1111/j.0006-341X.2000.00227.x> extended to time-varying covariates. Estimation and variable selection are based on the methodology described in Beretta and Heuchenne (2019) <doi:10.1080/02664763.2018.1554627>; confidence intervals of the parameter estimates may be computed using a bootstrap approach. Moreover, data following the PH cure model may be simulated using a method similar to Hendry (2014) <doi:10.1002/sim.5945>, where the event-times are generated on a continuous scale from a piecewise exponential distribution conditional on time-varying covariates.

Alessandro Beretta

penPHcure

Variable Selection in PH Cure Model with Time-Varying Covariates

C<c3><a9>dric Heuchenne

penPHcure.object function

a list with elements named <code>CURE</code> and <code>SURV</code>, containing the penalty weights. For more details, see <code><a rd-options='' href='penPHcure'>penPHcure</a></code>.

character string indicating the method used to transform the covariates included in the incidence (cure) component from time-varying to time-invariant. See <code><a rd-options='' href='penPHcure'>penPHcure</a></code> for more details.

This class of objects is returned by the function <code><a rd-options='' href='penPHcure'>penPHcure</a></code> when is called with the argument <code>pen.type = "SCAD" | "LASSO"</code>. Objects of this class have methods for the functions <code>summary</code> and <code>predict</code>.

penPHcure.object: Penalized PH cure model object

Description

Arguments

Details

See Also