Paper Key : IRJ************393
Author: Pratibha Singh,Ram Krishna Paramhans Dubey
Date Published: 04 Jan 2025
Abstract
Distributed Denial-of-Service (DDoS) attacks pose a significant threat to network security, disrupting services by overwhelming systems with a flood of malicious traffic. The detection and mitigation of such attacks require robust machine learning techniques capable of handling large volumes of data and adapting to the evolving nature of attack strategies. This paper presents a comprehensive review of semi-supervised machine learning techniques for DDoS attack detection and mitigation. Semi-supervised learning, which leverages both labeled and unlabeled data, offers an efficient solution for DDoS detection, especially in scenarios where labeled attack data is scarce or costly to obtain. We explore various algorithms within the semi-supervised paradigm, including self-training, co-training, and graph-based methods, and their application to DDoS detection. The review also discusses the integration of semi-supervised learning with other advanced techniques such as anomaly detection, feature extraction, and deep learning to enhance detection accuracy and reduce false positives. Furthermore, the paper examines the challenges and opportunities presented by semi-supervised approaches, including data imbalance, real-time detection, and adaptive mitigation strategies. By analyzing the strengths and limitations of existing methods, this review aims to provide a foundation for future research and development in the field of DDoS attack defense using semi-supervised machine learning.