MULTI-STAGE ADVERSARIAL DEFENSE FOR ONLINE DDOS ATTACK DETECTION SYSTEM IN IOT

The rapid expansion of Internet of Things (IoT) devices has significantly increased the vulnerability of IoT networks to Distributed Denial of Service (DDoS) attacks. This project introduces a Multi-Stage Adversarial Defense for Online DDoS Attack Detection System, combining both machine learning and deep learning techniques to provide an effective and robust defense mechanism. The system utilizes a hybrid approach, integrating Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNN), LightGBM (Light Gradient Boosting Machine), and Naive Bayes for real-time attack detection and classification. Deep learning models like LSTM and CNN are employed for their ability to capture complex temporal and spatial patterns in IoT network traffic, while LightGBM and Naive Bayes, both traditional machine learning algorithms, offer efficient decision-making with structured data. The system categorizes traffic into five classes: Mirai, DoS (Denial of Service), Scan, Normal, and MITM (Man-in-the-Middle) ARP Spoofing. By leveraging the strengths of both machine learning and deep learning models, the system enhances accuracy, speed, and resilience, providing a comprehensive solution for mitigating DDoS attacks and safeguarding IoT networks against evolving threats.

Keywords: Multi-Stage Adversarial Defense, DDoS Attack Detection, IoT Security, LSTM, CNN, LightGBM, Naive Bayes, Machine Learning, Deep Learning, Real-Time Detection, Mirai Botnet, Denial of Service (DoS), MITM ARP Spoofing, Attack Classification, Cybersecurity.