EUVIMEDEUVIMED entdecken
Crossref

Towards Autonomous Network Security: An AI-Based Comparative Framework for Intelligent Traffic Analysis and Threat Detection

Senan A.Abd

Dijlah Journal of Engineering Sciences ISSN: 3078-9664, e-ISSN: 3078-9656 · 17.06.2026 · 3 (2)

DOI: 10.65204/djes.v3i2.713 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Dijlah Journal of Engineering Sciences ISSN: 3078-9664, e-ISSN: 3078-9656, online erschienen am 17.06.2026. Verfasst von Senan A.Abd.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this research, we compare the performance of traditional machine learning algorithms with a CNN-LSTM deep learning model for detecting intrusions into a network. We conducted our research using the NSL-KDD dataset, which is commonly used for evaluating the effectiveness of various intrusion detection methods. We evaluated seven traditional classifiers and a hybrid CNN-LSTM model, which captures both spatial and temporal patterns in network traffic. All models were trained using the KDDTrain subset and tested using the KDDTest subset. The KDDTest subset contains 17 types of attacks not present in KDDTrain to simulate the reality of zero-day vulnerability exploitation. The hybrid CNN-LSTM model produced the best classification results, achieving an overall accuracy rate of 81.21%, which was just slightly higher than the decision tree classifier at 80.90%. All models performed much worse (15-20% less accurately) when tested against attacks that had not been seen including tired methodologies, illustrating the difficulty in identifying new forms of attack. Attack classes with relatively low numbers of examples in the training set (such as R2L and U2R) exhibited very poor performance. These findings highlight the need for testing on actual unknown attacks to evaluate system performance; additionally, more sophisticated algorithms (like CNN-LSTMs) do not demonstrate significant advantages (compared to less sophisticated methods) when it comes to autonomous detection systems for protecting networks.

Zitieren

APA

A.Abd, S. (2026). Towards Autonomous Network Security: An AI-Based Comparative Framework for Intelligent Traffic Analysis and Threat Detection. Dijlah Journal of Engineering Sciences ISSN: 3078-9664, e-ISSN: 3078-9656, 3(2). https://doi.org/10.65204/djes.v3i2.713

Vancouver

A.Abd S. Towards Autonomous Network Security: An AI-Based Comparative Framework for Intelligent Traffic Analysis and Threat Detection. Dijlah Journal of Engineering Sciences ISSN: 3078-9664, e-ISSN: 3078-9656. 2026;3(2). doi:10.65204/djes.v3i2.713

BibTeX
@article{AAbd2026,
  author = {A.Abd, Senan},
  title = {Towards Autonomous Network Security: An AI-Based Comparative Framework for Intelligent Traffic Analysis and Threat Detection},
  journal = {Dijlah Journal of Engineering Sciences ISSN: 3078-9664, e-ISSN: 3078-9656},
  year = {2026},
  volume = {3},
  number = {2},
  doi = {10.65204/djes.v3i2.713}
}
Kostenlos starten →