Deep learning for intrusion detection in IoT networks

The rapid proliferation of Internet of Things (IoT) devices has transformed our daily lives, introducing innovations like smart homes, wearables, and advanced industrial automation. While these interconnected systems offer convenience and efficiency, they also present significant security challenges. The increasing complexity and scale of IoT networks elevate the risk of malicious attacks, making the protection of these networks a pressing concern. Despite advancements in intrusion detection systems (IDS), existing methods often struggle to balance accuracy and computational efficiency, leading to higher false positive rates and inadequate detection of sophisticated threats. Traditional machine learning approaches frequently fall short in adapting to the evolving nature of attacks, thereby highlighting the necessity for more robust solutions. In response to these challenges, our study proposes an ensemble learning approach that utilizes bagging to enhance intrusion detection systems in IoT environments. By integrating the strengths of Deep Neural Networks (DNNs) and Convolutional Neural Networks (CNNs), we aim to leverage their unique advantages to improve the detection accuracy and resilience of IDS against a diverse range of cyber threats. To evaluate the effectiveness of our proposed model, we utilized the Edge-IIoTset dataset, which contains a comprehensive range of network traffic scenarios, including both normal and malicious behaviors. This dataset enables us to rigorously train and assess our models, ensuring their applicability in real-world IoT security contexts.