AI in Network Security Enhancing DDoS Attack Detection and Mitigation through Machine Learning and Deep Learning

Abstract

The increase in popularity of the Distributed Denial of Service (DDoS) attacks poses significant problems for network security. The purpose of the present research is to explore how Artificial Intelligence (AI) contributes to the prevention and detection of these attacks using machine learning and deep learning platforms. In particular, the Support Vector Machine (SVM), Long Short-Term Memory (LSTM), and Random Forest (RF) models are used to determine their effectiveness in detecting and distinguishing healthy and malicious network traffic. The evaluation of accuracy, precision, recall, and F1-score of the employed models is determined through training and testing the two widely familiar datasets, CIC-DDoS2019 and NSL-KDD. The results revealed that SVM and LSTM models have promising results with a high level of precision score, but without a recall feature, which is important for DDOS attacks. However, the evidence strengthens the idea of using AI-supported applications in real-time threat detection and alleviation. This is mainly because they are more responsive than the current methods. Future research must improve the memory of models and test combined forms of AI to improve DDoS defense mechanisms.