fracridge

Is an implementation of fractional ridge regression (FRR).

Installation:

MATLAB

Download and copy the files from https://github.com/nrdg/fracridge/tree/master/matlab into your MATLAB path.

Python

To install the release version:

pip install fracridge

Or to install the development version (and stay up-to-date by pulling from GitHub):

pip install -e .

To upgrade to newer release version:

pip install -U fracridge

Usage

MATLAB

[coef,alphas] = fracridge(X,fracs,y,tol,mode)

Python

There's a functional API:

from fracridge import fracridge
coefs, alphas = fracridge(X, y, fracs)

Or a sklearn-compatible OO API:

from fracridge import FracRidge
fr = FracRridge(fracs=fracs)
fr.fit(X, y)
coefs = fr.coef_
alphas = fr.alpha_

Online documentation

https://nrdg.github.io/fracridge/

How to cite

If you use fracridge, please cite our paper: "Fractional ridge regression: a fast, interpretable reparameterization of ridge regression" (2020) GigaScience, Volume 9, Issue 12, December 2020, https://doi.org/10.1093/gigascience/giaa133 link.

For your convenience, here is the bibtex entry

@ARTICLE{fracridge2020,
  title    = "Fractional ridge regression: a fast, interpretable
              reparameterization of ridge regression",
  author   = "Rokem, Ariel and Kay, Kendrick",
  journal  = "Gigascience",
  volume   =  9,
  number   =  12,
  month    =  nov,
  year     =  2020
  }