MORE-PLR

This repository contains the source code for the paper "MORE–PLR: Multi-Output Regression Employed for Partial Label Ranking". The data shown in the paper can be found in the data/ folder.
If you are interested of applying these models yourself, take a look at the SimpleStart.py file. But generally the models were implemented in the typical sklearn style, with fit and predict.

Reproducibility

Firstly make sure to have scikit-lr installed (the current state-of-the-art). To install execute the following command in the scikit-lr folder: pip install --verbose --no-build-isolation --editable .; python setup.py build_ext --inplace General python requirements can be installed via enviroment.yml.
At the end one might need to execute python setup.py build_ext --inplace in the root folder, to build the Cython _overlap_intervals.pyx file.
The scripts used for the experiments can be found in MORE/experiments \

Results

Nicely structured you can find the presented results split up into three view options:

• data/results contains all the results for the evaluations as .csv file.
• data/tables contains the results represented as tables.
• data/reports contains the results as .html overview.

Cite

@inproceedings{thies2024more,
  title={MORE--PLR: Multi-Output Regression Employed for Partial Label Ranking},
  author={Thies, Santo MAR and Alfaro, Juan C and Bengs, Viktor},
  booktitle={International Conference on Discovery Science},
  pages={401--416},
  year={2024},
  organization={Springer}
}

Contact

For indepth questions on MORE feel free to contact S.Thies@campus.lmu.de or JuanCarlos.Alfaro@uclm.es.