Enhanced detection of zero-day vulnerabilities in network traffic using hybrid random forest and convolutional neural network model
Nature Journal of Emerging Sciences Technologies and Innovations · 28.01.2026 · 4 (3) · 205-220
DOI: 10.65752/7hxrth09 ↗Kurz gefasst
Fachartikel in der Zeitschrift Nature Journal of Emerging Sciences Technologies and Innovations, online erschienen am 28.01.2026. Verfasst von Fortune Daberechi Ifeanyi und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Zero-day vulnerabilities represent critical security threats that exploit previously unknown software flaws before patches become available. This study presents a novel hybrid machine learning approach combining Random Forest (RF) and Convolutional Neural Network (CNN) for enhanced zero-day vulnerability detection in network traffic, leveraging the CSE-CIC-IDS2018 dataset. The methodology employs rigorous preprocessing, including feature selection, label encoding, and standardisation. The RF classifier performs initial probabilistic classification, generating confidence scores that serve as input to the CNN for refined behavioral analysis. The integrated model achieved 95.37% accuracy with strong performance across precision (0.9331), recall (0.9337), and F1-score (0.9407). Comparative analysis with existing anomaly-based approaches demonstrates significant improvements in false-positive reduction and classification reliability. The system exhibited exceptional performance in detecting brute-force attacks (AUC = 0.9997) while maintaining robust detection capabilities for web-based threats. Results highlight the potential of supervised hybrid learning models for early detection of emerging cyber threats, offering a scalable solution for real-world network security applications.
Zitieren
APA
Ifeanyi, F. D., James, O. A., Abdullahi, S. M., & Eze, V. C. (2026). Enhanced detection of zero-day vulnerabilities in network traffic using hybrid random forest and convolutional neural network model. Nature Journal of Emerging Sciences Technologies and Innovations, 4(3), 205-220. https://doi.org/10.65752/7hxrth09
Vancouver
Ifeanyi FD, James OA, Abdullahi SM, Eze VC. Enhanced detection of zero-day vulnerabilities in network traffic using hybrid random forest and convolutional neural network model. Nature Journal of Emerging Sciences Technologies and Innovations. 2026;4(3):205-220. doi:10.65752/7hxrth09
BibTeX
@article{Ifeanyi2026,
author = {Ifeanyi, Fortune Daberechi and James, Onah Ahinyohe and Abdullahi, Samaila Musa and Eze, Victor Chidera},
title = {Enhanced detection of zero-day vulnerabilities in network traffic using hybrid random forest and convolutional neural network model},
journal = {Nature Journal of Emerging Sciences Technologies and Innovations},
year = {2026},
volume = {4},
number = {3},
pages = {205--220},
doi = {10.65752/7hxrth09}
}
EUVIMED