NEWS
SHAPBoost 1.0.3 (2026-01-19)
Bug Fixes
- Fixed a bug where R² metric was not being correctly maximized in SHAPBoostRegressor.
SHAPBoost 1.0.2 (2026-01-13)
New Features
- Added
fixed_variables parameter to SHAPBoostRegressor and SHAPBoostSurvival
to allow users to specify variables that should always be included in the selected feature set.
SHAPBoost 1.0.1 (2025-12-04)
Bug Fixes
- Fixed a bug that caused unsupported types in xgboost for SHAPBoostRegressor
- Updated compatibility with latest xgboost versions
SHAPBoost 1.0.0 (2025-09-29)
New Features
- Initial release of SHAPBoost package
- Implementation of SHAPBoost algorithm for feature selection using SHAP values
- Support for regression analysis with
SHAPBoostRegressor class
- Support for survival analysis with
SHAPBoostSurvival class
- Integration with XGBoost for gradient boosting
- Support for Cox proportional hazards models in survival analysis
- Cross-validation based feature selection
- Collinearity detection and handling
- Configurable metrics (MAE, MSE, R², C-index)
Dependencies
- Depends on R >= 3.5.0
- Imports: xgboost, SHAPforxgboost, methods, caret, Matrix
Documentation
- Complete documentation for all exported functions and classes
- Examples for both regression and survival analysis use cases