SENTINELGPT
International Journal of Service Science, Management, Engineering, and Technology · 11.11.2026 · 17 (1) · 1-23
DOI: 10.4018/ijssmet.427559 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Service Science, Management, Engineering, and Technology, erschienen am 11.11.2026. Verfasst von Nachaat Mohamed und 1 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 3.0).
Abstract (Originalsprache)
Security operation centers (SOCs) are overwhelmed by alert volumes, leaving analysts little time for strategic threat investigation. This paper presents SENTINELGPT, an autonomous large language model-agent framework for automated alert triage, threat hunting, and incident response in artificial intelligence-augmented SOC environments. Evaluated on the UNSW-NB15 dataset (257,673 records, 10 attack categories), the system achieves 90.19% accuracy, a 92.30% F1-score, an area under the curve of 0.9861, and a false positive rate of 1.76% using random forest. Multi-class triage yields a weighted F1-score of 72.39%, confirming that large language model-driven SOC agents can significantly reduce analyst workload while maintaining high detection fidelity.
Zitieren
APA
Mohamed, N., & Taherdoost, H. (2026). SENTINELGPT. International Journal of Service Science, Management, Engineering, and Technology, 17(1), 1-23. https://doi.org/10.4018/ijssmet.427559
Vancouver
Mohamed N, Taherdoost H. SENTINELGPT. International Journal of Service Science, Management, Engineering, and Technology. 2026;17(1):1-23. doi:10.4018/ijssmet.427559
BibTeX
@article{Mohamed2026,
author = {Mohamed, Nachaat and Taherdoost, Hamed},
title = {SENTINELGPT},
journal = {International Journal of Service Science, Management, Engineering, and Technology},
year = {2026},
volume = {17},
number = {1},
pages = {1--23},
doi = {10.4018/ijssmet.427559}
}
EUVIMED