EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

DESIGN AND EVALUATION OF A SCALABLE AI-POWERED BIG DATA ANALYTICS FRAMEWORK FOR EARLY CYBERSECURITY THREAT DETECTION

Waqar Azeem, Minahil Ashraf, Aatif Hussain, Hamna Ijaz, Madiha Ijaz, Muhammad Usman Ahmad

Pakistan Journal of Scientific Research · 28.03.2026 · 6 (01) · 77-84

DOI: 10.57041/gbn9h465 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Pakistan Journal of Scientific Research, online erschienen am 28.03.2026. Verfasst von Waqar Azeem und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The fast growth of cyber physical systems has made traditional cybersecurity tools less efficient. Most of them are unable to deal with complex and quick changing threats. This paper presents how AI-powered big data analysis could help in finding cyber threats on the early stages. By combining large-scale data processing with advanced machine learning methods this can be made possible. The proposed framework of this study uses a multi-layers system model for collecting processing and analysising data across multiple systems. AI and machine learning based analysis can handle real-time threat detection, and automatic alert and response actions. The method was tested by using three known datasets: ICIDS2017, UNSW-NB15, and NSL-KDD. These include different types of attacks including DoS, port scanning, botnet, and exploitation. The performance of three machine learning models Random Forest, XGBoost, and Support Vector Machine (SVM) was measured by using five performance metrics, accuracy, precision, recall, F1-score, and ROC-AUC. The XGBoost model performed best, with an accuracy of 99.14%, F1-score 98.97%, and the ROC AUC of 0.9982. It shows a beating performance compared to SVM. Random Forest also showed strong results, with accuracy of 98.76%. In this study a detailed review of recent progress, and challenges were also discussed that can assist future research in federated learning, explainable AI (XAI), and automatic threat response.

Zitieren

APA

Azeem, W., Ashraf, M., Hussain, A., Ijaz, H., Ijaz, M., & Ahmad, M. U. (2026). DESIGN AND EVALUATION OF A SCALABLE AI-POWERED BIG DATA ANALYTICS FRAMEWORK FOR EARLY CYBERSECURITY THREAT DETECTION. Pakistan Journal of Scientific Research, 6(01), 77-84. https://doi.org/10.57041/gbn9h465

Vancouver

Azeem W, Ashraf M, Hussain A, Ijaz H, Ijaz M, Ahmad MU. DESIGN AND EVALUATION OF A SCALABLE AI-POWERED BIG DATA ANALYTICS FRAMEWORK FOR EARLY CYBERSECURITY THREAT DETECTION. Pakistan Journal of Scientific Research. 2026;6(01):77-84. doi:10.57041/gbn9h465

BibTeX
@article{Azeem2026,
  author = {Azeem, Waqar and Ashraf, Minahil and Hussain, Aatif and Ijaz, Hamna and Ijaz, Madiha and Ahmad, Muhammad Usman},
  title = {DESIGN AND EVALUATION OF A SCALABLE AI-POWERED BIG DATA ANALYTICS FRAMEWORK FOR EARLY CYBERSECURITY THREAT DETECTION},
  journal = {Pakistan Journal of Scientific Research},
  year = {2026},
  volume = {6},
  number = {01},
  pages = {77--84},
  doi = {10.57041/gbn9h465}
}
Kostenlos starten →