summary.penPHcure {penPHcure}R Documentation

Summary method for penPHcure.object

Description

Produces a summary of a fitted penalized PH cure model, after selection of the tuning parameters, based on AIC or BIC criteria.

Usage

## S3 method for class 'penPHcure'
summary(object,crit.type=c("BIC","AIC"),...)

Arguments

object

an object of class penPHcure.object.

crit.type

a character string indicating the criterion used to select the tuning parameters, either "AIC" or "BIC". By default crit.type = "BIC".

...

ellipsis to pass extra arguments.

Value

An object of class summary.penPHcure, a list including the following elements:

N

the sample size (number of individuals).

censoring

the proportion of censored individuals.

K

the number of unique failure times.

isTies

logical value: TRUE in case of tied event times.

pen.type

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

crit.type

a character string indicating the criterion used to select tuning parameters, either "AIC" or "BIC". By default crit.type = "BIC".

tune_params

a list with elements named CURE and SURV containing the selected tuning parameters, which minimize the AIC/BIC criterion.

crit

value of the minimized AIC/BIC criterion.

CURE

a matrix where in the first column the estimated regression coefficients in the cure (incidence) component are provided. If the argument inference (in the penPHcure function) was set equal to TRUE, two additional columns for the confidence intervals are provided.

SURV

a matrix where in the first column the estimated regression coefficients in the survival (latency) component are provided. If the argument inference (in the penPHcure function) was set equal to TRUE, two additional columns for the confidence intervals are provided.

Examples

# Generate some data (for more details type ?penPHcure.simulate in your console)
set.seed(12) # For reproducibility
data <- penPHcure.simulate(N=250)

### Tune penalized cure model with SCAD penalties
# First define the grid of possible values for the tuning parameters.
pen.tuneGrid <- list(CURE = list(lambda = c(0.01,0.03,0.05,0.07,0.09),
                                 a = 3.7),
                     SURV = list(lambda = c(0.01,0.03,0.05,0.07,0.09),
                                 a = 3.7))
# Tune the penalty parameters.
tuneSCAD <- penPHcure(Surv(time = tstart,time2 = tstop,
                           event = status) ~ z.1 + z.2 + z.3 + z.4,
                      cureform = ~ x.1 + x.2 + x.3 + x.4,
                      data = data,pen.type = "SCAD",
                      pen.tuneGrid = pen.tuneGrid,
                      print.details = FALSE)
# Use the summary method to see the results
summary(tuneSCAD)
# 
# ------------------------------------------------------
# +++   PH cure model with time-varying covariates   +++
# +++             [ Variable selection ]             +++
# ------------------------------------------------------
# Sample size:  250
# Censoring proportion:  0.5
# Number of unique event times: 125
# Tied failure times:  FALSE
# Penalty type:  SCAD
# Selection criterion:  BIC
# 
# ------------------------------------------------------
# +++               Tuning parameters                +++
# ------------------------------------------------------
#  Cure (incidence) --- lambda:  0.07 
#                            a:  3.7 
# 
#  Survival (latency) - lambda:  0.07 
#                            a:  3.7 
# 
#  BIC =  -118.9359 
# 
# ------------------------------------------------------
# +++                Cure (incidence)                +++
# +++     [ Coefficients of selected covariates ]    +++
# ------------------------------------------------------
#              Estimate
# (Intercept)  0.872374
# x.1         -0.958260
# x.3          0.685916
# 
# ------------------------------------------------------
# +++              Survival (latency)                +++
# +++     [ Coefficients of selected covariates ]    +++
# ------------------------------------------------------
#      Estimate
# z.1  0.991754
# z.3 -1.008180

# By default, the summary method for the penPHcure.object returns the selected 
#  variables based on the BIC criterion. For AIC, the user can set the 
#  argument crit.type equal to "AIC". 
summary(tuneSCAD,crit.type = "AIC")
# 
# ------------------------------------------------------
# +++   PH cure model with time-varying covariates   +++
# +++             [ Variable selection ]             +++
# ------------------------------------------------------
# Sample size:  250
# Censoring proportion:  0.5
# Number of unique event times: 125
# Tied failure times:  FALSE
# Penalty type:  SCAD
# Selection criterion:  AIC
# 
# ------------------------------------------------------
# +++               Tuning parameters                +++
# ------------------------------------------------------
#  Cure (incidence) --- lambda:  0.07 
#                            a:  3.7 
# 
#  Survival (latency) - lambda:  0.07 
#                            a:  3.7 
# 
#  AIC =  -136.5432 
# 
# ------------------------------------------------------
# +++                Cure (incidence)                +++
# +++     [ Coefficients of selected covariates ]    +++
# ------------------------------------------------------
#              Estimate
# (Intercept)  0.872374
# x.1         -0.958260
# x.3          0.685916
# 
# ------------------------------------------------------
# +++              Survival (latency)                +++
# +++     [ Coefficients of selected covariates ]    +++
# ------------------------------------------------------
#      Estimate
# z.1  0.991754
# z.3 -1.008180


[Package penPHcure version 1.0.2 Index]