RymaBmzz / ExplainMultilabelClassifiers

MARLENA - Explain Multilabel Classifiers

MARLENA: Explain Multilabel Classifiers

This project aims to address the multi-label black-box outcome explanation problem. Introducing MARLENA (Multi-label Rule-based ExplaNAtions)!

@inproceedings{panigutti2019explaining,
  title={Explaining multi-label black-box classifiers for health applications},
  author={Panigutti, Cecilia and Guidotti, Riccardo and Monreale, Anna and Pedreschi, Dino},
  booktitle={International Workshop on Health Intelligence},
  pages={97--110},
  year={2019},
  organization={Springer}
}

You can find MARLENA python library here.

Running MARLENA

0.Install the requirements.txt listed packages

1. Install multilabelexplanations: In order to run MARLENA you first have to locally install the pyhton module multilabelexplanations. You can do this by running the following command into the module directory:

$ cd multilabelexplanations
$ pip install .

2. Run the python scripts: you have to run the code python folde in the listed order:
   1_prepare_datasets.py
   2_train_blackbox.py
   3_pairwise_distances.py
   4_global_decision_tree.py

3. Run MARLENA experiments py script the script to run the experiments is in the python folder

$ python experiments_tuned_optimized.py [dataset_name] [black_box_name]

4. Analyze the results using the Analysis jupyter notebooks

Outputs

You can find experiments results into the output folder.