mlr3cmprsk

This package is under development

Package website: release

mlr3cmprsk extends the mlr3 ecosystem with a unified interface for machine learning in competing risks survival analysis. It provides consistent task, learner, and prediction abstractions, enabling seamless benchmarking, model comparison, and integration with the broader mlr3 framework.

Installation

Install the development version from GitHub:

# install.packages("pak")
pak::pak("mlr-org/mlr3cmprsk")

Example

library(mlr3cmprsk)
set.seed(42)

task = tsk("pbc")
task$select(c("age", "chol", "albumin", "bili"))
task$set_col_roles(cols = "status", add_to = "stratum")

learners = lrns(c("cmprsk.fg", "cmprsk.aalen"))

bm_grid = benchmark_grid(task, learners, rsmp("cv", folds = 3))
bm = benchmark(bm_grid)

# AUC at t = 100 (mean over causes, weighted by event frequencies)
bm$score(msr("cmprsk.auc", time = 100))[, .(task_id, learner_id, iteration, cmprsk.auc)]

##    task_id   learner_id iteration cmprsk.auc
## 1:     pbc    cmprsk.fg         1  0.7813865
## 2:     pbc    cmprsk.fg         2  0.7663073
## 3:     pbc    cmprsk.fg         3  0.8390308
## 4:     pbc cmprsk.aalen         1  0.5000000
## 5:     pbc cmprsk.aalen         2  0.5000000
## 6:     pbc cmprsk.aalen         3  0.5000000

# Brier score at t = 100 (mean over causes, weighted by event frequencies)
bm$score(msr("cmprsk.brier", time = 100))[, .(task_id, learner_id, iteration, cmprsk.brier)]

##    task_id   learner_id iteration cmprsk.brier
## 1:     pbc    cmprsk.fg         1    0.1564736
## 2:     pbc    cmprsk.fg         2    0.1750273
## 3:     pbc    cmprsk.fg         3    0.1371346
## 4:     pbc cmprsk.aalen         1    0.2182071
## 5:     pbc cmprsk.aalen         2    0.2200391
## 6:     pbc cmprsk.aalen         3    0.2181846

For more competing risk learners, see the available list at mlr3extralearners.

For more details about available measures and their parameters, see reference list.

Code of Conduct

Please note that the mlr3cmprsk project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.