R package HybridFS

Description

This R package features:

1. HFS, an efficient nonlinear feature screening method using multiple utilities.
2. HiFIT, a high-dimensional feature importance testing via machine learning models.

Minimal Requirement

• System requirement: Rtools (Windows), None (Linux, Mac)

• Dependent on R (>= 4.0.0), methods, e1071, glmnet, randomForest, xgboost, foreach, parallel, doParallel, dHSIC, isotree

• Multi-core CPU for parallel computing

Installation

install.packages("devtools")
devtools::install_github("IV012/HybridFS")

Related Pages

• Simulation Studies and Real Data Analysis using HybridFS

• Deep Treatment Learning

References

Mi, X., Zou, B., Zou, F. et al. Permutation-based identification of important biomarkers for complex diseases via machine learning models. Nat Commun 12, 3008 (2021). https://doi.org/10.1038/s41467-021-22756-2