EAGLET-python

Combining multi-label classifiers based on projections of the output space using Evolutionary algorithms - based on a paper by Jose M. Moyano, Eva L. Gibaja, Krzysztof J. Cios, Sebastián Ventura

• Original Paper: https://doi.org/10.1016/j.knosys.2020.105770
• Original Source Code: Github

Disclaimer

This software is written as a student project and there is no warranty for it to work and/or return correct results.

This work is done by:

• A. Tohidi
• A. Foroutan

How to launch EAGLET-python?

Using a python virtual environment (venv) is suggested!

First, make sure wheel is installed in current environment:

pip install wheel

Then install requirements:

pip install -r "requirements.txt"

Note: use --default-timeout=1000 flag if you face a http timeout error.

Now you can run RunExperiment.py with config file path as an argument. For example:

Unix:

python3 RunExperiment.py ./Configs/emotions_config.json

Windows:

python RunExperiment.py ./Configs/emotions_config.json

Implementation References:

1. Datasets
2. Paper results
3. MLC concepts
4. arff file extension
5. F-Measure
6. Scikit-learn user guide
7. Scikit-multilearn user guide
8. scipy Sparse Matrix