Package: oscar 1.2.1

oscar: Optimal Subset Cardinality Regression (OSCAR) Models Using the L0-Pseudonorm

Optimal Subset Cardinality Regression (OSCAR) models offer regularized linear regression using the L0-pseudonorm, conventionally known as the number of non-zero coefficients. The package estimates an optimal subset of features using the L0-penalization via cross-validation, bootstrapping and visual diagnostics. Effective Fortran implementations are offered along the package for finding optima for the DC-decomposition, which is used for transforming the discrete L0-regularized optimization problem into a continuous non-convex optimization task. These optimization modules include DBDC ('Double Bundle method for nonsmooth DC optimization' as described in Joki et al. (2018) <doi:10.1137/16M1115733>) and LMBM ('Limited Memory Bundle Method for large-scale nonsmooth optimization' as in Haarala et al. (2004) <doi:10.1080/10556780410001689225>). The OSCAR models are comprehensively exemplified in Halkola et al. (2023) <doi:10.1371/journal.pcbi.1010333>). Multiple regression model families are supported: Cox, logistic, and Gaussian.

Authors:Teemu Daniel Laajala [aut, cre], Kaisa Joki [aut], Anni Halkola [aut]

oscar_1.2.1.tar.gz
oscar_1.2.1.zip(r-4.5)oscar_1.2.1.zip(r-4.4)oscar_1.2.1.zip(r-4.3)
oscar_1.2.1.tgz(r-4.4-x86_64)oscar_1.2.1.tgz(r-4.4-arm64)oscar_1.2.1.tgz(r-4.3-x86_64)oscar_1.2.1.tgz(r-4.3-arm64)
oscar_1.2.1.tar.gz(r-4.5-noble)oscar_1.2.1.tar.gz(r-4.4-noble)
oscar_1.2.1.tgz(r-4.4-emscripten)
oscar.pdf |oscar.html
oscar/json (API)
NEWS

# Install 'oscar' in R:
install.packages('oscar', repos = c('https://syksy.r-universe.dev', 'https://cloud.r-project.org'))

Peer review:

Bug tracker:https://github.com/syksy/oscar/issues

Uses libs:
  • fortran– Runtime library for GNU Fortran applications
Datasets:
  • ex_K - Example data from TYKS / HUSLAB
  • ex_X - Example data from TYKS / HUSLAB
  • ex_Y - Example data from TYKS / HUSLAB
  • ex_c - Example data from TYKS / HUSLAB

On CRAN:

fortranpenalized-regressionregression

4.34 score 2 stars 11 scripts 193 downloads 23 exports 7 dependencies

Last updated 1 years agofrom:7b4f61bcc8. Checks:OK: 4 NOTE: 5. Indexed: yes.

TargetResultDate
Doc / VignettesOKNov 21 2024
R-4.5-win-x86_64NOTENov 21 2024
R-4.5-linux-x86_64NOTENov 21 2024
R-4.4-win-x86_64NOTENov 21 2024
R-4.4-mac-x86_64NOTENov 21 2024
R-4.4-mac-aarch64NOTENov 21 2024
R-4.3-win-x86_64OKNov 21 2024
R-4.3-mac-x86_64OKNov 21 2024
R-4.3-mac-aarch64OKNov 21 2024

Exports:coefcostfeatkitsoscaroscar.binarizeoscar.binplotoscar.bsoscar.bs.boxplotoscar.bs.koscar.bs.plotoscar.bs.visuoscar.controloscar.cost.afteroscar.cvoscar.cv.visuoscar.paretooscar.pareto.visuoscar.sparsifyoscar.visuplotpredictshow

Dependencies:hamletlatticeMatrixplyrpROCRcppsurvival

Example use of the OSCAR package

Rendered fromexample.Rmdusingknitr::rmarkdownon Nov 21 2024.

Last update: 2023-08-28
Started: 2020-11-04

Readme and manuals

Help Manual

Help pageTopics
oscar: Optimal Subset Cardinality Regressionoscar-package
Extract coefficients of oscar-objectscoef,oscar-method plot,oscar-method predict,oscar-method
Return total cost of model fit based on provided kit/variable costs vectorcost cost,oscar-method cost,oscar-methods
Example data from TYKS / HUSLABex_c ex_K ex_X ex_Y
Return named vector of feature indices with a given k that are non-zerofeat feat,oscar-method feat,oscar-methods
Return named vector of indices for kits with a given k that are non-zerokits kits,oscar-method kits,oscar-methods
Main OSCAR fitting functionoscar
S4-class for oscaroscar-class
Binary logical indicator matrix representation of an oscar object's coefficients (zero vs. non-zero, i.e. feature inclusion)oscar.binarize
Visualize binary indicator matrix optionally coupled with cross-validation performance for oscar modelsoscar.binplot
Bootstrapping for oscar-fitted model objectsoscar.bs
Bootstrap visualization with boxplot, percentage of new additionsoscar.bs.boxplot
Reformatting bootstrap output for cardinality k rowsoscar.bs.k
Bootstrap heatmap plot for oscar modelsoscar.bs.plot
Visualize bootstrapping of a fit oscar objectoscar.bs.visu
Control OSCAR optimizer parametersoscar.control
Return total cost of model fits if the cost is not included in the oscar objectoscar.cost.after
Cross-validation for oscar-fitted model objects over k-rangeoscar.cv
Visualize cross-validation as a function of koscar.cv.visu
Retrieve a set of pareto-optimal points for an oscar-model based on model goodness-of-fit or cross-validationoscar.pareto
Visualize oscar model pareto frontoscar.pareto.visu
Create a sparse matrix representation of betas as a function of koscar.sparsify
Target function value and total kit cost as a function of number of kits includedoscar.visu
Showing oscar-objectsshow,oscar-method