Linear and logistic ridge regression functions. Additionally includes special functions for genome-wide single-nucleotide polymorphism (SNP) data.
The ridge package offers linear and logistic ridge regression, especially for small data sets and genome-wide SNP data.
If you found a bug or have suggestions, feel free to get in contact via steffen.moritz10 at gmail.com
Since I only have taken over this package (it was orphaned) - it might be that I can't help with all details.
Be aware, that therefore I also can not guarantee to 100% that everything works as expected.
Erika Cule published also a scientific paper about the method, which you can find here: https://onlinelibrary.wiley.com/doi/abs/10.1002/gepi.21750
Cule, Erika, and Maria De Iorio. "Ridge regression in prediction problems: automatic choice of the ridge parameter." Genetic epidemiology 37.7 (2013): 704-714.
bugfix in logisticRidge and linearRidge see: https://github.com/SteffenMoritz/ridge/issues/2
bugfix to pass cran checks 'warning: missing template: HAVE_GSL_HEADER autoheader'
Some fixed to remain on CRAN and comly with CRAN policy
Improved Description file
Changed NEWS file to markup document
Made package CRAN ready again
Created github repository for the package
Fixed warning using Wpendantic gcc
Adapted the DESCRIPTION file to latest CRAN requirements
Renamed CHANGELOG to NEWS
NEWS template update
Changes to .C Method registration
* Fixed layout of .Rd files * Added deletion of Makevars to cleanup script
* Flat text (.txt) data files were moved from ridge/data to ridge/inst/extdata (in the source, which becomes ridge/extdata in the installed package). The .txt files should be in inst/extdata because they files are used by the package examples (albeit in not run sections), as described in Writing R Extensions 1.1.5 Data in packages. * Some users were reporting problems installing the package on some Linux OS. configure has been modified to fix this problem.
* Bug fix in src/commonFunctions.c
* Added configure.ac so that package will install if GSL >= 1.14 is not available (with linearRidgeGenotypes, logisticRidgeGenotypes, linearRidgeGenotypesPredict and logisticRidgeGenotypesPredict disabled). * configure.ac detects whether openblas is available and if it is found, links to that. This speeds up computation. (http://xianyi.github.com/OpenBLAS/)
fixed a bug in linearRidge when scaling = "none"
added functions linearRidgeGenotypes and logisticRidgeGenotypes and their predicting counterparts linearRidgeGenotypesPredict and logisticRidgeGenotypespredict. These functions fit linear and logistic ridge regression models for genome-wide SNP data, optionally automatically choosing the ridge parameter