Repository navigation
spfcICOMP 0.1.0
This is the first CRAN release of spfcICOMP, an R package for shrinkage principal fitted components with information complexity-based model selection in high-dimensional regression and classification.
Main capabilities
- Regularised covariance estimation using Oracle Approximating Shrinkage and Maximum Entropy Covariance
- Response-guided feature screening for high-dimensional predictors
- Structural-dimension selection using conventional and information-complexity criteria
- Reduced-space regression and classification
- Train-test splitting and cross-validation utilities
- Simulation tools for assessing dimension recovery, feature screening and predictive performance
Installation
Install the released version directly from CRAN:
install.packages("spfcICOMP")
library(spfcICOMP)Release verification
Version 0.1.0 completed the following checks:
- 149 of 149 test expectations passed
- GitHub Actions checks passed on macOS, Windows, Ubuntu R-release, Ubuntu R-devel and Ubuntu R-oldrel
- Compatibility with the declared requirement of R 3.5.0 or later was audited
- CRAN incoming pretests completed with no errors or warnings
- The package was published on CRAN on 3 September 2026
Links
- CRAN package: https://cran.r-project.org/package=spfcICOMP
- Documentation: https://cran.r-project.org/web/packages/spfcICOMP/
- Reference manual: https://cran.r-project.org/web/packages/spfcICOMP/spfcICOMP.pdf
- Source repository: https://github.com/ilovemaths/spfcICOMP
- Issue tracker: https://github.com/ilovemaths/spfcICOMP/issues
Citation
Please use the package citation supplied by R:
citation("spfcICOMP")The release translates methodological work on shrinkage covariance estimation, sufficient dimension reduction and information-complexity-based model selection into documented and reproducible statistical software.