spect: Survival Prediction Ensemble Classification Tool

A tool for survival analysis using a discrete time approach with ensemble binary classification. 'spect' provides a simple interface consistent with commonly used R data analysis packages, such as 'caret', a variety of parameter options to help facilitate search automation, a high degree of transparency to the end-user - all intermediate data sets and parameters are made available for further analysis and useful, out-of-the-box visualizations of model performance. Methods for transforming survival data into discrete-time are adapted from the 'autosurv' package by Suresh et al., (2022) <doi:10.1186/s12874-022-01679-6>.

Version: 1.0
Depends: R (≥ 4.0), futile.logger, dplyr
Imports: doParallel, ggplot2, survminer, riskRegression, caret, caretEnsemble, survival, rlang
Suggests: knitr, rmarkdown, randomForest, kernlab, testthat (≥ 3.0.0)
Published: 2025-04-08
DOI: 10.32614/CRAN.package.spect
Author: Stephen Abrams [aut, cre]
Maintainer: Stephen Abrams <stephen.abrams at gmail.com>
BugReports: https://github.com/dawdawdo/spect/issues
License: GPL-3
URL: https://github.com/dawdawdo/spect
NeedsCompilation: no
Materials: README
CRAN checks: spect results

Documentation:

Reference manual: spect.pdf
Vignettes: create_synthetic_data (source, R code)
modeling_with_binary_classifiers (source, R code)
training_synthetic_data (source, R code)

Downloads:

Package source: spect_1.0.tar.gz
Windows binaries: r-devel: spect_1.0.zip, r-release: spect_1.0.zip, r-oldrel: not available
macOS binaries: r-devel (arm64): spect_1.0.tgz, r-release (arm64): spect_1.0.tgz, r-oldrel (arm64): spect_1.0.tgz, r-devel (x86_64): spect_1.0.tgz, r-release (x86_64): spect_1.0.tgz, r-oldrel (x86_64): spect_1.0.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=spect to link to this page.