MSGL README
In msgl: Multinomial Sparse Group Lasso

set.seed(150)
pkg_version <- packageVersion("msgl")

Multinomial sparse group lasso

Multiclass classification with feature and parameter selection using sparse group lasso for the multinomial model. Suitable for high dimensional problems.

This is the R package msgl version r pkg_version.

R-package Overview

This package implements procedures for working with multinomial logistic regression models using sparse group lasso. This includes procedures for fitting and cross validating sparse models in a high dimensional setup. See the Getting started with msgl (predict primary cancer site based on microRNA measurements) for an example of a workflow consisting of 1) model selection and assessment using cross validation, 2) estimation of a final model and 3) using the selected model for carrying out predictions on new data.

alt tag

Classification of cancer site. Error estimated by 10-fold cross validation on a data set consisting of microRNA expression measurements of laser dissected primary cancers.

Package highlights:

Feature and parameter selection
Fast coordinate gradient descent algorithm
Suitable for high dimensional multiclass classification
Support for lasso, group lasso and sparse group lasso
Supports custom grouping of features
Supports sample weighting
Supports individual weighting of the group and parameter penalties

The penalized maximum likelihood estimator for multinomial logistic regression is computed using a coordinate gradient descent algorithm via the sglOptim optimizer. Use of parallel computing for cross validation and subsampling is supported through the foreach and doParallel packages.

Installation

Install the released version from CRAN:

install.packages("msgl")

Install the version from GitHub:

# install.packages("devtools")
devtools::install_github("nielsrhansen/sglOptim", build_vignettes = TRUE)
devtools::install_github("nielsrhansen/msgl", build_vignettes = TRUE)

If you don't want to build the vignettes when installing, just remove the build_vignettes = TRUE argument.

Minimal Example

library(msgl)

# Load some data
data(PrimaryCancers)

# Setup 2 parallel units
cl <- makeCluster(2)
registerDoParallel(cl)

# Do 10-fold cross validation on 100 models with increasing complexity, using the 2 parallel units
fit.cv <- msgl::cv(
  x = x,
  classes = classes,
  alpha = 0.5,
  lambda = 0.5,
  use_parallel = TRUE
)

stopCluster(cl)

# Print information about models
# and cross validation errors
fit.cv