How to setup FakeNewsDetector

pip install -r requirements.txt

python -m spacy download en

python feature_generator.py

Note on the data:

• The data were borrowed from another researcher and thus are not hosted on the public Github repo. Please contact us and we can request permission to share the data.

Purpose of each file:

test_validation.py:

• Creates and runs the entire model, including both logistic regressions, the CNN and the RNN
• This includes training the models and evaluating them on the final test set

feature_characteristics.tsv

• Contains statistics on how the various features vary across the different veracity labels

feature_generator.py

• Contains functions for making features across the dataset (which is a collection XML files)

model.ipynb

• Notebook containing a quick runthrough of our model, including some notes on how we implemented the various steps.
• Mostly used for testing purposes, this is not meant to be a full running version of the model. See test_validation.py for full model

parser.py

• Contains functions for reading through the XML data, writing in feature values and getting statistics on feature values