Implementations of different NLP methods towards classifying text for the purpose of detecting fake news as part of the course "IN4325 - Information Retrieval" of TUDelft's masters program.
Using FakeNewsNet and the included GossipCop and PolitiFact dataset.
Train models to classify data using different methods:
- Tf-Idf
- Bag-Of-Words
- Doc2Vec (DBOW/DDM/combined)
- GloVe
- BERT
Report and compare results for above methods using:
- Accuracy
- F1 score