srlearn

srlearn copied to clipboard

######## srlearn ########

.. image:: https://raw.githubusercontent.com/srlearn/srlearn/main/docs/source/_static/preview.png :alt: Repository preview image: "srlearn. Python wrappers around BoostSRL with a scikit-learn-style interface. pip install srlearn."

|License|_ |LGTM|_ |GitHubBuilds|_ |Codecov|_ |ReadTheDocs|_

.. |License| image:: https://img.shields.io/github/license/srlearn/srlearn.svg :alt: License .. _License: LICENSE

.. |LGTM| image:: https://img.shields.io/lgtm/grade/python/github/srlearn/srlearn?label=code%20quality&logo=lgtm :alt: LGTM code quality analysis .. _LGTM: https://lgtm.com/projects/g/srlearn/srlearn/context:python

.. |GitHubBuilds| image:: https://github.com/srlearn/srlearn/actions/workflows/python_tests.yml/badge.svg :alt: GitHub CI Builds .. _GitHubBuilds: https://github.com/srlearn/srlearn/actions/workflows/python_tests.yml

.. |Codecov| image:: https://codecov.io/gh/srlearn/srlearn/branch/main/graphs/badge.svg?branch=main :alt: Code coverage status .. _Codecov: https://codecov.io/github/srlearn/srlearn?branch=main

.. |ReadTheDocs| image:: https://readthedocs.org/projects/srlearn/badge/?version=latest :alt: Documentation status .. _ReadTheDocs: https://srlearn.readthedocs.io/en/latest/

srlearn is a Python package for learning statistical relational models, and wraps BoostSRL <https://starling.utdallas.edu/software/BoostSRL>_ (and other implementations <https://github.com/srlearn/SRLBoost>_) with a scikit-learn interface.

• Documentation: https://srlearn.readthedocs.io/en/latest/
• Questions? Contact Alexander L. Hayes <https://hayesall.com>_ (hayesall <https://github.com/hayesall>_)

Getting Started

Prerequisites:

• Java (1.8, 1.11)
• Python (3.7, 3.8, 3.9, 3.10)

Installation

.. code-block:: bash

pip install srlearn

Basic Usage

The general setup should be similar to scikit-learn. But there are a few extra requirements in terms of setting background knowledge and formatting the data.

A minimal working example (using the Toy-Cancer data set imported with 'load_toy_cancer') is:

.. code-block:: python

from srlearn.rdn import BoostedRDNClassifier
from srlearn import Background
from srlearn.datasets import load_toy_cancer
train, test = load_toy_cancer()
bk = Background(modes=train.modes)
clf = BoostedRDNClassifier(
    background=bk,
    target='cancer',
)
clf.fit(train)
clf.predict_proba(test)
# array([0.88079619, 0.88079619, 0.88079619, 0.3075821 , 0.3075821 ])
print(clf.classes_)
# array([1., 1., 1., 0., 0.])

train and test are each srlearn.Database objects, so this hides some of the complexity behind the scenes.

This example abstracts away some complexity in exchange for compactness. For more examples, see the Example Gallery <https://srlearn.readthedocs.io/en/latest/auto_examples/index.html>_.

Citing

If you find this helpful in your work, please consider citing:

.. code-block:: bibtex

@misc{hayes2019srlearn, title={srlearn: A Python Library for Gradient-Boosted Statistical Relational Models}, author={Alexander L. Hayes}, year={2019}, eprint={1912.08198}, archivePrefix={arXiv}, primaryClass={cs.LG} }

Contributing

Many thanks to those who have already made contributions:

• Alexander L. Hayes <https://hayesall.com>_, Indiana University, Bloomington
• Harsha Kokel <https://harshakokel.com/>_, The University of Texas at Dallas
• Siwen Yan <https://dtrycode.github.io/>_, The University of Texas at Dallas

Many thanks to the known and unknown contributors to WILL/BoostSRL/SRLBoost, including: Navdeep Kaur, Nandini Ramanan, Srijita Das, Mayukh Das, Kaushik Roy, Devendra Singh Dhami, Shuo Yang, Phillip Odom, Tushar Khot, Gautam Kunapuli, Sriraam Natarajan, Trevor Walker, and Jude W. Shavlik.

We have adopted the Contributor Covenant Code of Conduct <https://github.com/srlearn/srlearn/blob/main/.github/CODE_OF_CONDUCT.md>_ version 1.4. Please read, follow, and report any incidents which violate this.

Questions, Issues, and Pull Requests are welcome. Please refer to CONTRIBUTING.md <https://github.com/srlearn/srlearn/blob/main/.github/CONTRIBUTING.md>_ for information on submitting issues and pull requests.

Versioning and Releases

We use SemVer <https://semver.org>_ for versioning. See Releases <https://github.com/srlearn/srlearn/releases>_ for stable versions that are available, or the Project Page on PyPi <https://pypi.org/project/srlearn/>_.