To run
to run and evaluate the model type 'python fnc_kfold_our_model.py' in console
This model is still under construction in order to achieve more boost in accuracy
to run and evaluate the baseline model run 'python fnc_kfold_their.py' in console after commenting out the classification section using the other classifiers (Since we have included more classifiers on the baseline).
Dataset will be available on request
In feature engineering the doc2vec training model is defined. It generates semantic features of the headline and body. In feature Engineering there is also another method named we have added to get the cosine similarity mesurement of the headline and body vectors
More information about the competition can be found on http://www.fakenewschallenge.org/
Our code is built upon the base line implementation described in https://github.com/FakeNewsChallenge/fnc-1-baseline