EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An Approach to Detect Man-in-the-Middle Attack in Internet of Things Using Semi-Supervised Classification Techniques

Devpriya Panda, Neelamadhab Padhy, Kavita Sharma, Prabodh K. Sahoo, Aswini Kumar Samantaray

AI, Computer Science and Robotics Technology · 21.07.2026 · 5

DOI: 10.5772/acrt.20250138 ↗

Kurz gefasst

Fachartikel in der Zeitschrift AI, Computer Science and Robotics Technology, online erschienen am 21.07.2026. Verfasst von Devpriya Panda und 4 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Man-in-the-middle (MitM) attacks pose a severe threat to Internet of Things (IoT) ecosystems by stealthily intercepting or altering communications between two parties who believe they are directly connected. These attacks enable unauthorized eavesdropping, data modification, and injection, leading to operational disruptions and breaches of privacy. To address this challenge, we propose a machine learning-based detection framework tailored for IoT environments using a semi-supervised learning approach. The model comprises two primary modules: the first handles data pre-processing, including outlier removal, normalization, and dimensionality reduction, while the second focuses on classification. The dataset considered comprises both normal and MitM attack traffic, with a mix of numerical and categorical features, ensuring realistic evaluation conditions. We use a 20:80 labelled–unlabelled data ratio to simulate real-world constraints. Three classifiers – label propagation, random forest with self-training, and semi-supervised support vector machines (S3VM) – are implemented and compared. Experimental evaluation on a benchmark IoT security dataset containing 25,000 records demonstrates that our model achieves a detection accuracy of 95.79% with S3VM, 93.1% with self-training random forest, and 91.6% with label propagation. Precision, recall, and F1-scores for all models exceed 90%. These results confirm the effectiveness of the proposed model in detecting MitM attacks even with limited labelled data. Furthermore, the high detection performance combined with low false positive rates and reduced computational overhead highlights its suitability for real-time deployment in resource-constrained IoT environments.

Zitieren

APA

Panda, D., Padhy, N., Sharma, K., Sahoo, P. K., & Samantaray, A. K. (2026). An Approach to Detect Man-in-the-Middle Attack in Internet of Things Using Semi-Supervised Classification Techniques. AI, Computer Science and Robotics Technology, 5. https://doi.org/10.5772/acrt.20250138

Vancouver

Panda D, Padhy N, Sharma K, Sahoo PK, Samantaray AK. An Approach to Detect Man-in-the-Middle Attack in Internet of Things Using Semi-Supervised Classification Techniques. AI, Computer Science and Robotics Technology. 2026;5. doi:10.5772/acrt.20250138

BibTeX
@article{Panda2026,
  author = {Panda, Devpriya and Padhy, Neelamadhab and Sharma, Kavita and Sahoo, Prabodh K. and Samantaray, Aswini Kumar},
  title = {An Approach to Detect Man-in-the-Middle Attack in Internet of Things Using Semi-Supervised Classification Techniques},
  journal = {AI, Computer Science and Robotics Technology},
  year = {2026},
  volume = {5},
  doi = {10.5772/acrt.20250138}
}
Kostenlos starten →