Version 0.2.0 - D-VARS and Robustness Improvements
This release introduces Derivative-based Global Sensitivity Measures (D-VARS) to the package, providing a powerful new tool for sensitivity analysis. It also includes a new, more robust implementation of the D-VARS method for improved accuracy and stability.
New Features
D-VARS Sensitivity Analysis: The dvars_sensitivities function has been added to perform D-VARS analysis.
Improvements
Robust D-VARS Implementation: A new dvars_sensitivities_robust function is now available. This is the recommended method for D-VARS analysis, as it uses a more advanced and stable approach for hyperparameter optimization.
Bug Fixes
Added joblib to the list of Python dependencies to resolve a missing dependency issue.
Fixed a bug in the D-VARS tests to ensure they run correctly.