HOW TO DETECT ANOMALIES IN NETWORK TRAFFIC USING RNN

Abstract

This study proposes a method for automatic anomaly detection using a recurrent neural network (LSTM RNN) based on network traffic metadata. The model examines temporal patterns of network flows and identifies deviations from normal situations as an attack. The results indicate that the model has high accuracy and stability.