SHAPBoost is an R package for the implementation of the SHAPBoost feature selection algorithm, which is a boosting method that uses SHAP values for feature ranking and selects in an iterative forward fashion. It is designed to work with regression and survival analysis.

Installation

You can install the development version of SHAPBoost from GitHub with:

# install.packages("pak")
pak::pak("O-T-O-Z/SHAPBoost-R")

Regression example

For regression tasks, SHAPBoost can be used with various evaluators such as linear regression or XGBoost (xgb). For metrics, it support mae (Mean Absolute Error), mse (Mean Squared Error), and r2 (R-squared or