Sorted L1 Penalized Estimation

Efficient implementations for Sorted L-One Penalized Estimation (SLOPE): generalized linear models regularized with the sorted L1-norm (Bogdan et al. 2015). Supported models include ordinary least-squares regression, binomial regression, multinomial regression, and Poisson regression. Both dense and sparse predictor matrices are supported. In addition, the package features predictor screening rules that enable fast and efficient solutions to high-dimensional problems.

SLOPE

Efficient implementations for Sorted L-One Penalized Estimation (SLOPE): generalized linear models regularized with the sorted L1-norm.

Features

• Gaussian (quadratic), binomial (logistic), multinomial logistic, and Poisson regression
• Sparse and dense input matrices
• Efficient hybrid coordinate descent algorithm
• Predictor (feature) screening rules that speed up fitting in high-dimensional settings
• Cross-validation
• Parallelized routines
• Duality-based stopping criteria for robust control of suboptimality

Installation

You can install the current stable release from CRAN with the following command:

install.packages("SLOPE")

Alternatively, you can install the development version from GitHub with the following command:

# install.packages("pak")
pak::pak("jolars/SLOPE")

Getting Started

By default, SLOPE fits a full regularization path to the given data. Here is an example of fitting a logistic SLOPE model to the built-in heart dataset.

library(SLOPE)

fit <- SLOPE(heart$x, heart$y, family = "binomial")

We can plot the resulting regularization path:

plot(fit)

We can also perform cross-validation to select optimal scaling of the regularization sequence:

set.seed(18)

cvfit <- cvSLOPE(heart$x, heart$y, family = "binomial")
plot(cvfit)

Ecosystem

SLOPE is also available as a

• Python package: sortedl1
• Julia package: SLOPE.jl
• C++ library: libslope

Versioning

SLOPE uses semantic versioning.

Code of conduct

Please note that the ‘SLOPE’ project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.