Designing An Intelligent Intrusion Detection System (Ids) Using Neural Networks To Detect Zero-Day Threats

Read full text Summarise with VIRAT
Authors: Qayssar Al-Omairi, Akram Hatem Saber, Haider Hadi Abbas, Munther Hameed Mohammed Ameen, Alaa Raad Salman
Cited by: 0
References: 14

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The ever-growing speed of developing the network infrastructures, along with an unparalleled growth of the load of traffic flows and their complexity, has made the systems vulnerable to the advanced cyber threats. Among such, zero-day attacks present one of the most severe threats to the security requirements of a particular cybersecurity environment, being based on new vulnerabilities that may resist the signature-based defense mechanisms. The conventional intrusion detection systems (IDS) are usually unable to scale, support high-dimensional traffic, and generalize thus proving inefficient in dynamic and real-life situations. In an attempt to overcome these issues, this research paper presents an ensemble-based intrusion detection model that will combine an autoencoder neural network and a Random Forest classifier. The suggested methodology is tested on large-scale, heterogeneous network traffic data which consists of millions of records constituting various types of attacks. results show that the ensemble model is almost optimal and has almost 99 percent accuracy, precision, recall and F1-score of known attacks. Furthermore, when training the model to zero-day attacks in which a specific set of attack types is omitted during the training process the model still has high recall (95-97%).

Fields of Study & Subjects

Network Security and Intrusion Detection Network Packet Processing and Optimization Artificial Immune Systems Applications

Publication details

Journal / source
International Journal of Advances in Signal and Image Sciences
Volume
12
Issue
3s
Pages
1315–1326
Publication date
9 March 2026
Language
EN