The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider removing the loading script and relying on automated data support (you can use convert_to_parquet from the datasets library). If this is not possible, please open a discussion for direct help.

Fake.br

Dataset Summary

Fake.Br Corpus is composed of aligned true and fake news written in Brazilian Portuguese.

Supported Tasks and Leaderboards

The task is text classification of news content.

Languages

The dataset is in Portuguese.

Citation Information

If you use "Fake.br Dataset", please include a citation to the project website and the corresponding paper published in PROPOR 2018 conference:

@InProceedings{fakebr:18,
  author={Monteiro, Rafael A. and Santos, Roney L. S. and Pardo, Thiago A. S. and de Almeida, Tiago A. and Ruiz, Evandro E. S. and Vale, Oto A.},
  title={Contributions to the Study of Fake News in Portuguese: New Corpus and Automatic Detection Results},
  booktitle={Computational Processing of the Portuguese Language},
  year={2018},
  publisher={Springer International Publishing},
  pages={324--334},
  isbn={978-3-319-99722-3},
}

or the paper published in Expert Systems with Applications:

@article{silva:20,
  title = "Towards automatically filtering fake news in Portuguese",
  journal = "Expert Systems with Applications",
  volume = "146",
  pages = "113199",
  year = "2020",
  issn = "0957-4174",
  doi = "https://doi.org/10.1016/j.eswa.2020.113199",
  url = "http://www.sciencedirect.com/science/article/pii/S0957417420300257",
  author = "Renato M. Silva and Roney L.S. Santos and Tiago A. Almeida and Thiago A.S. Pardo",
}

Contributions

Thanks to @ju-resplande for adding this dataset.

Downloads last month
46