Botnet Attack Detection in IoT Using a Hybrid Deep Learning Framework with ACLR Technique

Abstract

Cyber-attacks have increased significantly with the rapid expansion and widespread adoption of internet technologies, with botnet attacks emerging as one of the most destructive threats. The identification of botnets has become increasingly challenging due to diverse attack vectors and the continuous evolution of malware. In the context of the rapidly growing Internet of Things (IoT), many interconnected devices with limited security features are highly vulnerable to botnet intrusions, leading to substantial losses across various sectors. Traditional detection methods often fail to effectively identify such sophisticated and dynamic attack patterns. To address these challenges, this research proposes a hybrid machine learning model named ACLR, which integrates Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and Recurrent Neural Networks (RNN) in a stacked architecture. The model leverages the UNSW-NB15 dataset, comprising nine distinct attack categories, to learn complex patterns in network traffic data. Experimental results demonstrate that the proposed model achieves high detection accuracy and effectively identifies diverse botnet behaviours. Overall, the ACLR model enhances IoT security by providing a robust, scalable, and automated solution for detecting malicious network activities