EUVIMEDEUVIMED entdecken
Crossref

Machine Learning-Based Http Traffic Anomaly Identification in Internet of Things Environments

Ankit Kumar

Advances in Nonlinear Variational Inequalities · 04.03.2025 · 28 (6s) · 1058-1071

DOI: 10.52783/anvi.v28.7072 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Advances in Nonlinear Variational Inequalities, online erschienen am 04.03.2025. Verfasst von Ankit Kumar.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The rapid proliferation of Internet of Things (IoT) devices has fundamentally transformed computing paradigms, simultaneously creating unprecedented cybersecurity vulnerabilities. IoT devices communicate predominantly via HTTP, making HTTP traffic analysis a critical vector for anomaly detection.

Objective

This paper presents a comprehensive comparative study and novel framework for ML-based HTTP traffic anomaly identification in IoT environments, addressing unique challenges including resource constraints, heterogeneous device profiles, and real-time detection requirements.

Methods

We propose IoT-HTTPAD (IoT HTTP Traffic Anomaly Detection), a hybrid framework combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for spatial and temporal feature extraction. The framework was evaluated on four benchmark datasets: N-BaIoT, UNSW-NB15, TON-IoT, and CIC-IoT-2023, comprising over 13.8 million traffic records across diverse IoT attack scenarios.

Results

IoT-HTTPAD achieves a classification accuracy of 99.1%, precision of 98.5%, recall of 98.8%, and an AUC-ROC of 0.9960, outperforming seven baseline algorithms including Random Forest, XGBoost, SVM, and standalone LSTM. The framework demonstrates a 23.4% improvement in detection rate over traditional signature-based methods while maintaining a false positive rate below 0.9%.

Conclusion

The proposed IoT-HTTPAD framework provides a scalable, resource-aware solution for real-time HTTP anomaly detection in heterogeneous IoT ecosystems, with demonstrated effectiveness across DDoS, botnet infiltration, command injection, and man-in-the-middle attack categories.

Zitieren

APA

Kumar, A. (2025). Machine Learning-Based Http Traffic Anomaly Identification in Internet of Things Environments. Advances in Nonlinear Variational Inequalities, 28(6s), 1058-1071. https://doi.org/10.52783/anvi.v28.7072

Vancouver

Kumar A. Machine Learning-Based Http Traffic Anomaly Identification in Internet of Things Environments. Advances in Nonlinear Variational Inequalities. 2025;28(6s):1058-1071. doi:10.52783/anvi.v28.7072

BibTeX
@article{Kumar2025,
  author = {Kumar, Ankit},
  title = {Machine Learning-Based Http Traffic Anomaly Identification in Internet of Things Environments},
  journal = {Advances in Nonlinear Variational Inequalities},
  year = {2025},
  volume = {28},
  number = {6s},
  pages = {1058--1071},
  doi = {10.52783/anvi.v28.7072}
}
Kostenlos starten →