This project was inspired by the following paper: Anomoly Detection
This project takes real network traffic data which is annotated between malicious and not malicious and feeds that data into a bidirectional RNN to achieve an AUC of 0.83 while the paper achieves an AUC of 0.84
The Github: Link