Fake News Classifier

  • Built a Bi-LSTMs based system to identify unreliable news using a dataset containing 21000 news articles.
  • Performed standard text pre-processing and used word embeddings to boost test accuracy up to 95%.