v0.1.1

✨ New Features: Scikit-Learn Compatible Estimators

We are excited to introduce full scikit-learn compatibility! ReHLine now provides plq_Ridge_Classifier and plq_Ridge_Regressor estimators that integrate seamlessly with the entire scikit-learn ecosystem.

This means you can:

• Drop ReHLine estimators directly into your existing scikit-learn Pipeline.
• Perform robust hyperparameter tuning using GridSearchCV.
• Use standard scikit-learn evaluation metrics and cross-validation tools.

What's Changed

• update 'verbose' option by @Leona-LYT in #14
• add path solution and warm-start examples by @Leona-LYT in #15
• fix layout issue by @Leona-LYT in #16
• Update constraint.rst by @DataboyUsen in #17
• add CQR_Ridge_path_sol with examples by @Leona-LYT in #18
• modified _make_constraint_param by @Leona-LYT in #19
• modified sen_idx and add _sklearn_mixin.py by @Leona-LYT in #20
• delete _make_fair_classification function and modified related test file by @Leona-LYT in #22
• Update _base.py by @DataboyUsen in #21
• add documentation and tutorial example by @Leona-LYT in #23

Full Changelog: v0.1.0...v0.1.1

v0.1.0

What's Changed

• Overrides the get_params method in sklearn.BaseEstimator to fix the bug by @statmlben
• Update _base.py by @DataboyUsen in #11
• modified tiny error in _loss.py and add a path solution file by @Leona-LYT in #13
• Path solution v1 via warmstart in _path_sol.py by @Leona-LYT

New Contributors

• @DataboyUsen made their first contribution in #11
• @Leona-LYT made their first contribution in #13

Full Changelog: v0.0.6...v0.0.7

v0.0.7

What's Changed

• Overrides the get_params method in sklearn.BaseEstimator to fix the bug by @statmlben
• Update _base.py by @DataboyUsen in #11
• modified tiny error in _loss.py and add a path solution file by @Leona-LYT in #13
• Path solution v1 via warmstart in _path_sol.py by @Leona-LYT

New Contributors

• @DataboyUsen made their first contribution in #11
• @Leona-LYT made their first contribution in #13

Full Changelog: v0.0.6...v0.0.7

v0.0.6

What's Changed

• Add warmstart for the main solver by @statmlben in 48ae8b4
• Add composite quantile regressor by @youqiangao in #10

New Contributors

• @youqiangao made their first contribution in #10

Full Changelog: v0.0.5...v0.0.6

Release Note for v0.0.5

What's Changed

Major Changes:

• @yixuan Improved Building System:

  • Fixed the setup.py script to support building on MacOS environments.
  • The Github Action configuration can now automatically upload wheels to PyPI on new releases.

• @keepwith Add example of Rank Regression

  • Demo for using ReHLine solve Rank Regression

• @statmlben Tutorials for Loss and Constraints

  • add tutorials

Minor Changes:

• @statmlben
  • add convergence warning

New Contributors

Full Changelog: v0.0.4...v0.0.5

Release Note for v0.0.4

What's Changed

Major Changes:

• @statmlben Revised Classes:

  • ReHLine: This class is designed for manually setting up problems, providing users with greater flexibility and control.
  • plqERM_ridge: This class focuses on solving Empirical Risk Minimization (ERM) problems, enhancing the library's capabilities in this area.

• @statmlben Documentation Improvements:

  • Extensive enhancements have been made to the documentation, ensuring clearer guidance and support for users.

Minor Changes:

• add epsilon-insensitive loss for SVR by @Aoblex in #5
• Add a linear term to rehline.cpp by @aorazalin in #4

New Contributors

• @Aoblex made their first contribution in #5
• @aorazalin made their first contribution in #4

Full Changelog: v0.0.3...v0.0.4

Release Note for v0.0.3

Version 0.03

New Features

• Wheel Package Support for Linux and Windows:
  • Added pre-built wheel packages for Linux and Windows operating systems.
  • This simplifies the installation process and eliminates the need for manual compilation.

Thank you @JiantingFeng for contributing the GitHub Action file.

What's Changed

• test: set random seed and add default C by @JiantingFeng in #1
• Fix Windows Build Issue by Removing poetry build Command by @JiantingFeng in #2
• Chore and CI Enhancements by @JiantingFeng in #3

New Contributors

• @JiantingFeng made their first contribution in #1

Full Changelog: v0.1...v0.0.3

Release Note for v0.0.1

We are delighted to announce a new GitHub release of "ReHLine", a dedicated solver for Regularized Composite ReLU-ReHU Loss Minimization.