The adaptive firewall with log predictive analysis based on neural network
Abstract
The article considers the development of an adaptive IDS/IPS system based on neural networks, capable of detecting both known and previously unknown attacks. The analysis is conducted using the NF-ToN-IoT dataset. Three neural networks were trained: for attack detection, attack type identification, and unknown threat prediction. The results demonstrate high attack detection accuracy (96.97 %) and the ability to identify new threats (81.94 %), surpassing existing solutions.The developed firewall and intrusion prevention system demonstrates high efficiency, enabling the creation of a domestic security system capable of minimizing the risk of hacker attacks and ensuring reliable protection of network infrastructure.
Fields of Study & Subjects
Network Security and Intrusion Detection
Network Packet Processing and Optimization
Artificial Immune Systems Applications
Publication details
- Journal / source
- INFORMACIONNYE TEHNOLOGII
- Volume
- 32
- Issue
- 3
- Pages
- 149–156
- Publication date
- 13 March 2026
- Language
- EN