idsdarg/fakenewsdeeplearning

This is the code to reproduce the results in the paper:

"The Language of Fake News: Opening the Black-Box of Deep Learning Based Detectors", O'Brien et al. (2018)

Running the code from scratch

0. Pre-requisites: Python3 + packages: nltk, numpy, sklearn, Tensorflow (tested in version 1.12.0rc0)

1. Download and ungzip GoogleNews-vectors-negative300.bin.gz. Save the uncompressed GoogleNews-vectors-negative300.bin in the root directory of the repository (same directory as train.py). You can get the file here:

   https://github.com/mmihaltz/word2vec-GoogleNews-vectors

2. Run pattern removal script, clean_data.py:

python clean_data.py

3. Train the Neural Network: train.py (experiment could be either Trump or all)

python train.py --experiment=Trump

Stop the training when the validation accuracy does not increase anymore. The validation accuracy is displayed every 100 training steps. A directory in 'run' that cointains the network parameters is created.

4. Test the Neural Network eval.py

python eval.py --experiment=Trump

5. Get the most relevant patterns for each article:

python get_patterns.py --experiment=Trump

6. Display the most relevant patters accross all the dataset by parts of speech:

python parts_of_speech.py --experiment=Trump

7. We obtained the following patterns:

Frequent patterns useful in fake and real news

Dataset

The dataset consists on the Fake News Dataset by Kaggle collected by the BS detector (7,401 articles) + articles collected from "The Guardian" and "The New York Times" (8,999 articles).

In the data directory it can be found:

• data/raw: the original articles

• data/processed: the articles after removing words that are not in the English dictionary via PyEnchant

• data/clean: the articles after cleaning advertisements and announcements, punctuation, etc. This is the data before going to the detector.