Split Regularized Regression

Functions for computing split regularized estimators defined in Christidis, Lakshmanan, Smucler and Zamar (2019) . The approach fits linear regression models that split the set of covariates into groups. The optimal split of the variables into groups and the regularized estimation of the regression coefficients are performed by minimizing an objective function that encourages sparsity within each group and diversity among them. The estimated coefficients are then pooled together to form the final fit.

SplitReg

This package provides functions for computing the split regularized regression estimators defined in Christidis, Lakshmanan, Smucler and Zamar (2019).

You can install the stable version on R CRAN.

install.packages('SplitReg', dependencies = TRUE)

You can install the development version from GitHub

library(devtools)
devtools::install_github("AnthonyChristidis/SplitReg")

Usage

# A small example
library(MASS)
library(SplitReg)
set.seed(1)
beta <- c(rep(5, 5), rep(0, 45))
Sigma <- matrix(0.5, 50, 50)
diag(Sigma) <- 1
x <- mvrnorm(50, mu = rep(0, 50), Sigma = Sigma)
y <- x %*% beta + rnorm(50)
fit <- cv.SplitReg(x, y, num_models=10) # Use 10 models
coefs <- predict(fit, type="coefficients")

License

This package is free and open source software, licensed under GPL (>= 2).