Note: If you need the full corpus, please fill this form: MultiSourceFake data

This is the code of the system proposed in the paper:

FakeFlow: Fake News Detection by Modeling the Flow of Affective Information

REQUIREMENTS:

• gensim==3.8.0
• joblib==0.14.1
• Keras==2.2.4
• Keras-Preprocessing==1.1.0
• keras-self-attention==0.35.0
• numpy==1.16.0
• pandas==0.24.2
• nltk==3.4.5
• scikit-learn==0.20.2
• tensorflow-gpu==1.14.0
• tqdm==4.32.1
• hyperopt==0.1.1

Place your data in the folder ./data/DATASET_NAME

To run the model, run the file: fake_flow.py

parameters: -d: dataset name (i.e. MultiSourceFake).

-sn: number of segments.

-s: to search for params; enter a number larger than 0 to search for N different combination of parameters (e.g. 150).

-m: mode (train or test); if you want to load a pretrained model.

An example:

fake_flow.py -d MultiSourceFake -sn 10

To load saved model after training:

fake_flow.py -d MultiSourceFake -sn 10 -m test

To search for best params:

fake_flow.py -d MultiSourceFake -s 80

Citation:

@inproceedings{ghanem2021fakeflow,
  title={{FakeFlow: Fake News Detection by Modeling the Flow of Affective Information}},
  author={Ghanem, Bilal and Ponzetto, Simone Paolo and Rosso, Paolo and Rangel, Francisco},
  booktitle={Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics},
  year={2021}
}