Anomaly, attack, and vulnerability detection in transmitted network traffic using machine learning algorithms
Acta Polytechnica · 10.11.2026 · 66 (5) · 553-564
DOI: 10.14311/ap.2026.66.0553 ↗Kurz gefasst
Fachartikel in der Zeitschrift Acta Polytechnica, online erschienen am 10.11.2026. Verfasst von Erika Abigail Katonova und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This research focuses on developing of a network monitoring tool designed to detect vulnerabilities and suspicious patterns in the communication networks of two levels: application and standard user data flows. We demonstrate that the ML algorithm can improve the real-time detection of security threats within network infrastructures. A CNN-LSTM autoencoder mode was proposed, which learned to reconstruct network flows from normal data and recognise anomalies based on reconstruction errors. Also, a LSTM and Vulnerability classifier was proposed for further classification of attacks and vulnerabilities in anomalous data. The results of the tests demonstrated the model’s high success rate in detecting anomalous traffic. The proposed system was able to detect anomalies, four attack types (Portscan, SQL injection, XSS, and DoS) and three vulnerability classes (SQL, Redis, and sensitive data transmitted in cleartext). The F1 score for anomaly detection was 0.89 and the error rate for attack detection was in an interval of 1–5 %.
Zitieren
APA
Katonova, E. A., Petija, R., Fecilak, P., Goloviznina, T., & Cider, R. (2026). Anomaly, attack, and vulnerability detection in transmitted network traffic using machine learning algorithms. Acta Polytechnica, 66(5), 553-564. https://doi.org/10.14311/ap.2026.66.0553
Vancouver
Katonova EA, Petija R, Fecilak P, Goloviznina T, Cider R. Anomaly, attack, and vulnerability detection in transmitted network traffic using machine learning algorithms. Acta Polytechnica. 2026;66(5):553-564. doi:10.14311/ap.2026.66.0553
BibTeX
@article{Katonova2026,
author = {Katonova, Erika Abigail and Petija, Rastislav and Fecilak, Peter and Goloviznina, Tatiana and Cider, Radoslav},
title = {Anomaly, attack, and vulnerability detection in transmitted network traffic using machine learning algorithms},
journal = {Acta Polytechnica},
year = {2026},
volume = {66},
number = {5},
pages = {553--564},
doi = {10.14311/ap.2026.66.0553}
}
EUVIMED