Introduction

fastcmprsk is an R package for performing Fine-Gray regression via a forward-backward scan algorithm.

Official release is available on CRAN and the master branch on GitHub.

Features

• Scalable Fine-Gray estimation procedure for large-scale competing risks data.
• Currently supports unpenalized and penalized (LASSO, ridge, SCAD, MCP, elastic-net) regression.
• Can perform CIF estimation with interval/band estimation via bootstrap.

What’s New in Version 1.1.0?

• Official version is loaded onto CRAN.

Implementation

fastcmprsk in an R package with most functionality implemented in C. The package uses cyclic coordinate descent to optimize the likelihood function.

Installation

To install the latest development version, install from GitHub.

install.packages("devtools")
devtools::install_github(“erickawaguchi/fastcmprsk”)

System Requirements

Requires R (version 3.5.0 or higher).

User Documentation

• Package manual: Currently unavailable.
• Please cite Kawaguchi et al. (2019).

License

fastcmprsk is licensed under GPL-3.

Development

fastcmprsk is being developed in R Studio.