ENSC 833 Project

Performance Evaluation of Machine Learning Models Using IODA Data

Name: Mohid Hasan

Student ID: 301588534

Abstract

Internet anomalies may result from malware outbreaks, infrastructure failures, and conflict-driven connectivity disruptions. Early detection is essential for maintaining network stability, improving cybersecurity visibility, and ensuring resilience during crises. This project evaluates eight machine learning models for anomaly detection using Internet Outage Detection and Analysis (IODA) datasets within a sliding-window framework that captures temporal dependencies.

Two case studies, Iran and Cuba, are analyzed to represent both severe shutdowns and subtle connectivity degradations. The models are evaluated using ROC-AUC, PR-AUC, Accuracy, F1-score, confusion-matrix-based metrics, Precision, Sensitivity, and training time. Results show that temporal models such as RNN-GRU and RNN-Bi-LSTM provide strong and balanced detection performance, while Broad Learning System variants achieve near-top accuracy with significantly lower training cost, making them suitable for real-time deployment. The results also highlight that outage severity affects detection difficulty and that sliding-window modeling effectively captures temporal disruption patterns.

Final Deliverables

đź“„ Final Report (PDF) đź“Š Final Presentation (PPTX)

References

  1. Z. Li and L. Trajković, “Enhancing cyber defense,” Proc. IEEE SMC, 2023.
  2. L. F. Oliveira et al., “Internet outages during times of conflict,” Proc. IEEE SMC, 2024.
  3. Z. Li et al., “Machine learning for anomaly detection,” IEEE Journal on Selected Areas in Communications.
  4. CyberDefense GitHub. [Online]. Available: https://github.com/zhida-li/CyberDefense
  5. RouteViews. [Online]. Available: http://www.routeviews.org/
  6. Internet Outage Detection and Analysis (IODA). [Online]. Available: https://ioda.inetintel.cc.gatech.edu/
  7. L. F. Oliveira, “paper smc 2024” GitHub repository. [Online]. Available: https://github.com/luizsoliveira/paper_smc_2024
  8. M. Hasan, “anomaly detection” GitHub repository. [Online]. Available: https://github.com/mohid-hasan/anomaly_detection

Last updated: April 2026