Regression and Supervised Learning

Oct 13, 2025 by recoveryshake.com 35 views

hdcuremodels: Penalized Mixture Cure Models for High-Dimensional Data

Overview

The hdcuremodels package is a novel implementation of penalized mixture cure models for high-dimensional data. This package provides functions for fitting various parametric and semi-parametric mixture cure models with different penalty functions, testing for a significant cure fraction, and testing for sufficient follow-up. The package also includes false discovery rate controlled variable selection using model-X knock-offs.

Scope

The hdcuremodels package falls under the category of Regression and Supervised Learning.

Pre-submission Inquiry

No pre-submission inquiry has been approved for this package.

General Information

Target Audience and Scientific Applications

The target audience for this package is analysts who model time-to-event outcomes when some subjects either experience long-term survival or are not susceptible to the event of interest (simplistically, cured). The package is designed to provide a comprehensive solution for fitting and testing mixture cure models in high-dimensional data.

Novelty of Algorithm

The hdcuremodels package is the first implementation of penalized mixture cure models in R. The package provides an improvement on other implementations of similar algorithms in R by incorporating different penalty functions and false discovery rate controlled variable selection.

Badging

The package aims for a Silver badge, which requires a demonstrated generality of usage beyond one single envisioned use case. The package fulfils this aspect by providing a comprehensive solution for fitting and testing mixture cure models in high-dimensional data.

Technical Checks

The package has undergone the following technical checks:

The author has read the rOpenSci packaging guide.
The author has read the author guide and expects to maintain this package for at least 2 years or has another maintainer identified.
The author has read the Statistical Software Peer Review Guide for Authors.
The author has run autotest checks on the package, and ensured no tests fail.
The package does not violate the Terms of Service of any service it interacts with.
The package has a CRAN and OSI accepted license.
The package contains a README with instructions for installing the development version.

Publication Options

The author intends for this package to go on CRAN after the ROpenSci review process.

Code of Conduct

The author agrees to abide by rOpenSci's Code of Conduct during the review process and in maintaining the package should it be accepted.

Package Description

Package: hdcuremodels
Title: Penalized Mixture Cure Models for High-Dimensional Data
Version: 0.0.2
Date: 2025-03-11
Authors@R: 
    c(person("Han", "Fu", role = "aut"), person(c("Kellie J."), "Archer", email=
    "archer.43@osu.edu", role = c("aut","cre"), comment = c(ORCID="0000-0003-1555-5781")))
Description: Provides functions for fitting various penalized parametric and semi-parametric mixture cure models with different penalty functions, testing for a significant cure fraction, and testing for sufficient follow-up as described in Fu et al (2022)<doi:10.1002/sim.9513> and Archer et al (2024)<doi:10.1186/s13045-024-01553-6>. False discovery rate controlled variable selection is provided using model-X knock-offs. 
License: MIT + file LICENSE
Encoding: UTF-8
Depends: R (>= 4.2.0)
Imports: doParallel,
         flexsurv,
         flexsurvcure,
         foreach,
         ggplot2,
         ggpubr,
         glmnet,
         knockoff,
         mvnfast,
         parallel,
         plyr,
         methods,
         survival
Roxygen: list(markdown = TRUE, roclets = c ("namespace", "rd", "srr::srr_stats_roclet"))
RoxygenNote: 7.3.2
Suggests: 
    knitr,
    rmarkdown,
    roxygen2,
    testthat (>= 3.0.0)
VignetteBuilder: knitr
LazyData: true
URL: https://github.com/kelliejarcher/hdcuremodels
BugReports: https://github.com/kelliejarcher/hdcuremodels/issues
Config/testthat/edition: 3

Installation

You can install the hdcuremodels package from GitHub using the following code:

devtools::install_github("kelliejarcher/hdcuremodels")

Usage

The package provides a comprehensive solution for fitting and testing mixture cure models in high-dimensional data. The package includes functions for fitting various parametric and semi-parametric mixture cure models with different penalty functions, testing for a significant cure fraction, and testing for sufficient follow-up. The package also includes false discovery rate controlled variable selection using model-X knock-offs.

Example Use Cases

The package can be used in various applications, including:

Fitting mixture cure models in high-dimensional data
Testing for a significant cure fraction
Testing for sufficient follow-up
Performing false discovery rate controlled variable selection using model-X knock-offs

Conclusion

The hdcuremodels package provides a comprehensive solution for fitting and testing mixture cure models in high-dimensional data. The package includes functions for fitting various parametric and semi-parametric mixture cure models with different penalty functions, testing for a significant cure fraction, and testing for sufficient follow-up. The package also includes false discovery rate controlled variable selection using model-X knock-offs. The package aims for a Silver badge and has undergone the necessary technical checks.
hdcuremodels: Penalized Mixture Cure Models for High-Dimensional Data - Q&A

Overview

Q&A

Q: What is the main purpose of the `hdcuremodels` package?

A: The main purpose of the hdcuremodels package is to provide a comprehensive solution for fitting and testing mixture cure models in high-dimensional data.

Q: What types of mixture cure models are supported by the package?

A: The package supports various parametric and semi-parametric mixture cure models with different penalty functions.

Q: What is the significance of the cure fraction in mixture cure models?

A: The cure fraction is the proportion of individuals who are not susceptible to the event of interest and experience long-term survival.

Q: How does the package perform false discovery rate controlled variable selection?

A: The package uses model-X knock-offs to perform false discovery rate controlled variable selection.

Q: What are the benefits of using the `hdcuremodels` package?

A: The package provides a comprehensive solution for fitting and testing mixture cure models in high-dimensional data, which can be beneficial in various applications, including cancer research and epidemiology.

Q: Is the package suitable for users with limited programming experience?

A: The package is designed to be user-friendly, but users with limited programming experience may need to consult the documentation and examples provided.

Q: Can the package be used for other types of data besides high-dimensional data?

A: The package is specifically designed for high-dimensional data, but users may be able to adapt the package for other types of data with some modifications.

Q: How can users contribute to the development of the package?

A: Users can contribute to the development of the package by reporting bugs, suggesting new features, and providing feedback on the package's documentation and examples.

Q: Is the package compatible with other R packages?

A: The package is designed to be compatible with other R packages, but users may need to check the compatibility of the package with their specific use case.

Q: Can users request new features or functionality for the package?

A: Yes, users can request new features or functionality for the package by submitting an issue on the package's GitHub page.

Q: How can users get help with using the package?

A: Users can get help with using the package by consulting the documentation, examples, and online resources provided, as well as by submitting an issue on the package's GitHub page.

Example Use Cases

The package can be used in various applications, including:

Fitting mixture cure models in high-dimensional data
Testing for a significant cure fraction
Testing for sufficient follow-up
Performing false discovery rate controlled variable selection using model-X knock-offs

Conclusion

The hdcuremodels package provides a comprehensive solution for fitting and testing mixture cure models in high-dimensional data. The package includes functions for fitting various parametric and semi-parametric mixture cure models with different penalty functions, testing for a significant cure fraction, and testing for sufficient follow-up. The package also includes false discovery rate controlled variable selection using model-X knock-offs. Users can contribute to the development of the package by reporting bugs, suggesting new features, and providing feedback on the package's documentation and examples.

Installation

You can install the hdcuremodels package from GitHub using the following code:

devtools::install_github("kelliejarcher/hdcuremodels")

Usage

Example Code

# Load the package
library(hdcuremodels)

# Fit a mixture cure model
model <- fit_mixture_cure(data = my_data, penalty = "l1")

# Test for a significant cure fraction
test_cure_fraction(model = model)

# Test for sufficient follow-up
test_follow_up(model = model)

# Perform false discovery rate controlled variable selection
select_variables(model = model)

Note: The above code is just an example and may need to be modified to suit the specific use case.