GenAI-Based Threat Intelligence Model: A Comparative Performance Analysis with Traditional Tools
Küçük Resim Yok
Tarih
2025
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Institute of Electrical and Electronics Engineers Inc.
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
This study includes the architecture, modeling process and comparative performance analysis of a GenAI based large language model (LLM) supported threat intelligence system. The model was trained with a total of 2.8 million threat data obtained from sources such as MITRE ATT&CK, VirusTotal, CICIDS, PhishTank. In the experiments, the model demonstrated high performance by reaching 96.2% in overall accuracy, 94.5% in F1-score and 3.1% in false positive rate. In the analysis specific to attack types such as phishing, DDoS and malware, 95.2%, 96.8% and 93.7% accuracy rates were obtained, respectively. In comparisons with traditional tools such as Splunk, Snort and ELK Stack, the GenAI model provided an average of 6.8% higher accuracy; it especially provided 22% better success in zero-day attack detection. The system, which has an average prediction time of 342 ms, reveals that it is a strong candidate in real-time threat intelligence with its high accuracy and low false alarm rate. © 2025 IEEE.
Açıklama
10th International Conference on Computer Science and Engineering, UBMK 2025 -- 17 September 2025 through 21 September 2025 -- Istanbul -- 214243
Anahtar Kelimeler
Cyber Threat Intelligence, GenAI, Large Language Model, Performance Analysis, Splunk
Kaynak
International Conference on Computer Science and Engineering, UBMK
WoS Q Değeri
Scopus Q Değeri
N/A
Cilt
Sayı
2025
