Distributed Denial of Service Attack Detection using Machine Learning and Deep Learning Approach

Abstract

A Distributed Denial of Service (DDoS) attack can disrupt the availability of resources and safety of online services by flooding targeted networks or servers with excessive traffic. This paper presents a dual-model detection system that uses machine learning and deep learning techniques to tackle the complexity of DDoS attacks. The first model comprises of a stacked ensemble of K-Nearest Neighbors (KNN), Decision Tree Classifier, and Logistic Regression. By combining the predictive capabilities of these algorithms, this model increases detection accuracy while identifying a broad range of assault patterns and behaviors. The ensemble model improves the system's resilience and generalizability to DDoS attack type by combining several classifiers. In addition, the second model automatically recognizes complex patterns in network traffic data using deep learning. This methodology is intended to identify subtle, intricate DDoS assault indicators that conventional techniques could miss. It is especially well-suited for real-time detection in a dynamic network, as its deep architecture enables it to process big datasets efficiently.