Open Access Open Access  Restricted Access Subscription Access

Malicious Activity Detection In IOT

Annem Naga Leena, Sowmya J, Padmasree N

Abstract


The proliferation of Internet of Things (IoT) devices necessitates sophisticated intrusion detection systems (IDS) to counter escalating cyber threats. This comparative study investigates the efficacy of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) in identifying botnet activity within IoT network traffic, utilizing the Bot-IoT dataset. A crucial preprocessing step involved employing a Decision Tree-based feature selection method to isolate the most discriminative attributes. These features were then structured differently for each model: as 2D grids for the CNN (to extract spatial features) and as sequential inputs for the   RNN   (to   capture   temporal


dependencies).The experimental evaluation demonstrated that the CNN model, specifically when utilizing the ReLU activation function, achieved superior performance in detecting botnet intrusions. The CNN model recorded an exceptional F1-Score of 0.9994. This result strongly suggests that modeling the network traffic as a spatial feature representation is significantly more effective than focusing on temporal sequences for this specific task of IoT botnet detection.


Full Text:

PDF

References


Al-Asfoor, M., Al-Asfoor, M., & Al- Fayoumi, N. (2023). Deep learning-based detection of IoT botnet attacks: An exploration of residual networks. International Journal of Safety and Security Engineering, 13(4), 715–722. [2] Zhou, Z., Liu, C., & Chang, Y. (2022). TCN enhanced novel malicious traffic detection for IoT devices. Connection Science, 34(1), 1322– 1341. [3] Abusitta, A., Al-omari, M., & Abu-taleb, M. (2025). Smart deep learning model for enhanced IoT intrusion detection. Scientific Reports, 15, Article 20577. [4] Ahmed, A., Ali, A., Aamir, M., Khan, Z., & Habib, U. (n.d.). A novel ensemble method for enhancing Internet of Things device security against botnet attacks. Decision Analytics Journal, 8, Article 100307. [5] Al- Saidi, N. M. G., Al-karkhi, M. A. A., & AL- Masodi, H. H. (2022). An intelligent detection of malicious intrusions in IoT

based on machine learning and deep learning techniques. Computer Science Journal of Moldova, 30(3), 285-305. [6] Dib, E., Abderrahmane, M., Melainini, A., & Moukhlissa, S. (2023). Deep-learning based detection for cyber-attacks in IoT networks: A distributed attack detection framework. Journal of Network and Systems Management, 31, Article 33. [7] Ullah, I., & Mahmoud, Q. H. (2022). A two-level hybrid model for anomalous activity detection in IoT networks. IEEE Access, 10, 27073- 27087. [8] Hussain, F., Hussain, R., Hassan,

S. A., & Hossain, E. (2024). A survey on deep learning for security and privacy in the internet of things. ACM Computing Surveys, 56(3), 1-38. [9] Almseidin, M., Al- kasassbeh, M., & Kovacs, L. (2023). A transformer-based model for intrusion detection systems in IoT environments. Journal of Network and Computer

Applications, 219, Article 103714. [10] Soe,

Y. N., Mha, H. T., & Ahn, S. G. (2023). A

lightweight CNN-based network intrusion detection system for IoT networks. IEEE Access, 11, 98315- 98327


Refbacks

  • There are currently no refbacks.