The R package ‘survivalSL’ contains a variety of functions to construct a super learner in the presence of censored times-to-event and to evaluate its prognostic capacities. Several learners are proposed: proportional hazard (PH) regressions, penalized PH semi-parametric models, accelerated failure times (AFT) models, neural networks, random survival forests, etc.). We proposed also a variety of loss functions for the estimation of the weights (concordance index (Pencina and Uno versions), Brier score, area under the time-dependent ROC curve, negative binomial log-likelihood, etc.). S3 methods are included to evaluate the predictive capacities, as well as predicting survival curves from new observations.
# Simulate a training and validation samples
<- 500 # sample size for validation
n.valid <- 200 # sample size for training
n.learn <- n.valid + n.learn # overall sample size
n
<- 50 # maximum follow-up time
max.time
<- 0; sd.x <- 1 # normal distribution of the quantitative predictors
mean.x <- .5 # proportion of the binary predictors
proba.x
<- 2; b <- .05 # Weibull baseline distribution of the PH model
a <- c(log(1.8), log(1.8), log(1.3), 0, 0, 0) # regression coefficients
beta
# simulation of the training and validation samples
set.seed(123)
<- rnorm(n, mean.x, sd.x)
x1 <- rbinom(n, 1, proba.x)
x2 <- rbinom(n, 1, proba.x)
x3 <- rnorm(n, mean.x, sd.x)
x4 <- rbinom(n, 1, proba.x)
x5 <- rbinom(n, 1, proba.x)
x6 <- cbind(x1, x2, x3, x4, x5, x6) # matrix of the potential predictors
x
<- 1/b*((-exp(-1*(x %*% beta))*(log(1-runif(n, 0, 1))))**(1/a)) # time to event
times <- runif(n, min=0, max=max.time)
censoring
<- ifelse(times <= censoring, 1, 0) # event status
status <- ifelse(times <= censoring, times, censoring) # follow-up times
obs.times
<- cbind(obs.times, status, as.data.frame(x))
data
<- list(data[1:n.valid,], data[(n.valid+1):n,])
data.simul
# model estimation with default parameters and three learners
<-Surv(obs.times,status) ~ x1+x2+x3+x4+x5+x6
formula<- survivalSL(formula=formula,
slres methods=c("LIB_COXen", "LIB_AFTgamma", "LIB_PHexponential"),
metric="auc", data=data.simul[[1]], seed=123, optim.method="SANN")
# prognostic capacities from training sample
summary(slres, digits=3)
#> p_ci uno_ci auc bs ibs ribs bll ibll ribll ll
#> 0.677 0.677 0.719 0.21 0.095 0.107 0.609 0.314 0.354 -1310.989
# prognostic capacities from validation sample
summary(slres, newdata=data.simul[[2]], digits=3)
#> p_ci uno_ci auc bs ibs ribs bll ibll ribll ll
#> 0.722 0.726 0.784 0.192 0.093 0.098 0.57 0.312 0.333 -472.514
To install the latest release from CRAN:
install.packages("survivalSL")
To install the development version from GitHub:
::install_github("foucher-y/survivalSL") remotes
You can report any issues at this link.