Welcome to SuperSurv!
SuperSurv is designed to be a unified ecosystem for
machine learning and survival analysis. The framework integrates modern
ensemble learning techniques for right-censored data and is motivated by
recent advances in machine learning-based survival curve estimation
(Westling et al., 2024) and unified ensemble modeling for survival
analysis (Lyu et al., 2026).
However, installing every optional machine learning engine at once can take a long time and cause dependency conflicts on some operating systems.
To make your experience as smooth as possible, SuperSurv
uses a Modular Dependency Philosophy.
The core package is incredibly lightweight and installs in seconds. Heavy machine learning libraries (like XGBoost or Elastic Net) are only required when you explicitly ask to use them!
Step 1: Install the Core Package
You can install the development version of SuperSurv
directly from GitHub using the devtools or
remotes package:
# Install devtools if you don't have it
if (!requireNamespace("devtools", quietly = TRUE)) {
install.packages("devtools")
}
# Install SuperSurv from GitHub
devtools::install_github("yuelyu21/SuperSurv")Once installed, you can load the package and view all the available modeling and screening wrappers:
library(SuperSurv)
# See all 24 prediction models and 6 screening algorithms.
list_wrappers()Step 2: Install Base Learners (Optional but Recommended)
SuperSurv relies on external CRAN packages to run its
various base learners. If you try to run surv.rfsrc without
having the randomForestSRC package installed,
SuperSurv will gently pause and remind you to install
it.
If you want to unlock the full power of the package right now, you can copy and paste the following script to install the most commonly used machine learning and interpretability engines:
# List of highly recommended modeling engines
ml_packages <- c(
"survival", # Classical Cox models
"randomForestSRC", # Random Survival Forests
"ranger", # Fast Random Forests
"grf", # Generalized Random Forests
"xgboost", # Extreme Gradient Boosting
"glmnet", # Elastic Net & Penalized Regression
"mboost", # Component-Wise Cox Boosting
"flexsurv", # Flexible Parametric Survival Models
"rpart", # Decision Trees
"survex" # Time-Dependent XAI (Interpretability)
)
# Identify which ones you are missing
missing_pkgs <- ml_packages[!(ml_packages %in% installed.packages()[,"Package"])]
# Install the missing ones
if(length(missing_pkgs)) install.packages(missing_pkgs)Specialized Packages
A few wrappers require specialized packages that you might only need
for niche use cases: * surv.svm: Requires
survivalsvm * surv.gam: Requires
mgcv * surv.coxboost: Requires
CoxBoost * surv.deepsurv and
surv.deephit: Require the R packages
survivalmodels and reticulate, plus importable
Python modules torch, torchtuples, and
pycox. SuperSurv does not install or modify Python
environments automatically.
Step 3: You’re Ready!
Your environment is now completely set up. You are ready to build your first optimal survival ensemble!
👉 Click here to proceed to Tutorial 1: The SuperSurv Ensemble
References
Westling, T., Luedtke, A., Gilbert, P. B., & Carone, M.
(2024).
Inference for treatment-specific survival curves using machine
learning.
Journal of the American Statistical Association.
doi:10.1080/01621459.2023.2205060
Lyu, Y., Huang, X., Lin, S. H., & Li, Z. (2026).
SuperSurv: A Unified Framework for Machine Learning Ensembles in
Survival Analysis.
bioRxiv.
doi:10.64898/2026.03.11.711010