Neural Networks and Network Protocols to Predict Network Behavior
Haydar Teymourlouei, Vareva E. Harris · 2022 International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCME) · 2022
Accurately predicting network behavior is a crucial task for cyber security professionals in industry, military, government, and many other fields. Effective, consistent predictions will enable security teams to predict, detect, and eliminate network anomalies in real-time. This research uses machine learning with neural networks and real-time network protocol monitoring to predict network behavior. The UNSW-NB15 dataset is used for testing the methodology. The results show the accuracy of the deep learning solution reaches 95.6%. The research provides evidence to support neural networks and network protocols as effective tools for predicting network behavior.