SLOPE

R
Python
Julia
C++

An efficient and fully-featured package for Sorted L-One Penalized Estimation

Authors

Johan Larsson

Małgorzata Bogdan

Krystyna Grzesiak

Mathurin Massias

Jonas Wallin

Published

18 December 2025

Details

arXiv, arXiv:2511.02430

Links

 

SLOPE (Sorted L-One Penalized Estimation) is a regularization method for high-dimensional regression that controls the false discovery rate while performing variable selection. Unlike LASSO which uses a single penalty parameter, SLOPE uses a sequence of sorted penalties that provide better statistical properties for multiple testing scenarios.

This package provides efficient, production-ready implementations across R, Python, and Julia, with a unified C++ core for performance. It includes support for generalized linear models, sparse matrices, and modern optimization algorithms, making it suitable for large-scale statistical learning problems in genomics, econometrics, and machine learning.

Citation

BibTeX citation:
@misc{larsson2025,
  author = {Larsson, Johan and Bogdan, Malgorzata and Grzesiak, Krystyna
    and Massias, Mathurin and Wallin, Jonas},
  publisher = {arXiv},
  title = {Efficient Solvers for {SLOPE} in {R,} {Python,} {Julia,} and
    {C++}},
  number = {arXiv:2511.02430},
  date = {2025/11/04},
  url = {http://arxiv.org/abs/2511.02430},
  doi = {10.48550/arXiv.2511.02430},
  langid = {en}
}
For attribution, please cite this work as:
Larsson, Johan, Malgorzata Bogdan, Krystyna Grzesiak, Mathurin Massias, and Jonas Wallin. 2025–11AD. “Efficient Solvers for SLOPE in R, Python, Julia, and C++.” arXiv. https://doi.org/10.48550/arXiv.2511.02430.