Fake news classifier

Abstract

A fake news detector based on the LIAR-PLUS dataset.

Getting Started

Create a conda environment with python 3.6.8

conda create --name pytorch1.1 python=3.6.8
conda activate pytorch1.1

Install all dependencies using conda env create -f environment.yml.

cd fake-news-classifier
mkdir datasets/

Download the LIAR_PLUS dataset into datasets/LIAR_PLUS and the pretrained word vectors
wiki-news-300d-1M.vec from fasttext
into datasets/word_embeddings.

Training

Train the model (for binary classification)

python main.py --phase train --binary yes

For hex classification use --binary no

Testing

Similar to training, provide the phase and binary state options through the CLI.

python main.py --phase test --binary yes

References

• Where is your Evidence: Improving Fact-checking by Justification Modeling
• 2015, Conroy, Niall J., Victoria L. Rubin, and Yimin Chen.
  “Automatic deception detection: Methods for finding fake news.”
  Proceedings of the Association for Information Science and Technology 52, no. 1 (2015): 1-4.
• Kaggle - Toxic Comments EDA by jagangupta
• Kaggle - Spooky NLP Modelling by arthurtok
• Kaggle - Quora Insincere questions by ziliwang
• Awesome Fake News - GitHub