In this repository I showed how to detect fake news. The dataset is devided into 2 classes. Using RNN I made a classifier that can identify whether the news is fake or true. After cleaning the data I performed tokenization, and encoded the data. Then using RNN architecture and LSTM model I build this classifier.
The dataset is provided along with source codes. In the notebook all instructions are provide. Some theory, intuition behind RNN, LSTM and problems with Gradient Descent(vinishing gradient) is covered. I trained the model with only 2 eopchs. That gave more than 98% acc in val and test set.