spareg is an R package for building ensembles of predictive generalized linear models tailored for high-dimensional data. It leverages a powerful combination of variable screening and random projection techniques to efficiently tackle large-scale modeling problems.
- Efficient modeling with large predictor sets
- Modular S3 architecture for screening and projection techniques
- Competitive predictive performance with low computational cost
- Extensible framework for custom screening and projection methods
You can install the development version of spareg from GitHub:
# install.packages("devtools")
devtools::install_github("lauravana/spareg")The CRAN version can be installed with
install.packages("spareg")