A Fake News Classifier is a machine learning model that distinguishes between legitimate and fake news articles. In the mentioned study, the researchers used a combination of deep learning embedding and LSTM to build a binary classifier for fake news detection.

The deep learning embedding technique used in this study is BERT (Bidirectional Encoder Representations from Transformers), a powerful language model that can understand the context of words in a sentence. The BERT model's outputs were then connected to an LSTM (Long Short-Term Memory) layer, a type of recurrent neural network that can learn long-term dependencies in sequential data.

The researchers trained and evaluated the model on the FakeNewsNet dataset, which contains two sub-datasets, PolitiFact and GossipCop. They compared the performance of their proposed model with base classification models and found that the proposed model outperformed the vanilla pre-trained BERT model by 2.50% and 1.10% in accuracy on PolitiFact and GossipCop datasets, respectively.

In summary, the researchers used a combination of BERT and LSTM to build a binary classifier for fake news detection. The proposed model outperformed the vanilla pre-trained BERT model, demonstrating the effectiveness of using deep learning embedding and LSTM for fake news classification.

Source code