Metadata-Version: 2.1
Name: rank2plan
Version: 0.1.2
Summary: 
Author: Ryan Wang
Author-email: ryanxiaowang2001@gmail.com
Requires-Python: >=3.8,<3.11
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Requires-Dist: numpy (>=1.22.0,<2.0.0)
Requires-Dist: pulp (>=2.8.0,<3.0.0)
Description-Content-Type: text/markdown

# rank2plan

[![Tests](https://github.com/ryanxwang/rank2plan/actions/workflows/pytest.yml/badge.svg?branch=main)](https://github.com/ryanxwang/rank2plan/actions/workflows/pytest.yml)

Implementation of constraint generation and column generation for solving large
L1-RankSVMs with hinge loss (with pair-specific gaps) and sample weights. This
is based on the work by Dedieu et al (2022) on solving large L1-SVMs with hinge
loss. See `documents/theory.pdf` for how we extend their work. The "2plan" part
of the package name comes from the tool being used to learn heuristics for
planning.

## Installation

Install with

```bash
pip install rank2plan
```

The PyPI release only supports `3.8 <= python <= 3.10` for now.

## References

- A. Dedieu, R. Mazumder, and H. Wang. Solving l1-regularized svms and related
linear programs: Revisiting the effectiveness of column and constraint
generation. J. Mach. Learn. Res., 23:164:1–164:41, 2022. [[URL]](http://jmlr.org/papers/v23/19-104.html).

