Enhancing DDoS Attack Detection Using Machine Learning: A Framework with Feature Selection and Comparative Analysis of Algorithms

Abstract

Distributed Denial of Service (DDoS) attacks are an ever-present threat to network security and can make online services hard for users to access. Conventional detection methods often struggle to effectively counter new and sophisticated DDoS attacks. This research article aims to assess the effectiveness of several machine learning methods in detecting distributed denial-of-service (DDoS) attacks. The evaluation is conducted using the DDOS attack SDN dataset, which is sourced from Google's research dataset. Various algorithms, including Random Forest, Decision Tree, Naive Bayes, and Support Vector Machine (SVM), are used for the purpose of analyzing network traffic data and detecting abnormal patterns that may indicate DDoS attacks. Results indicate that the Random Forest algorithm achieves the highest accuracy rate of 99.4% in detecting DDoS attacks. Additionally, the Decision Tree and SVM algorithms perform admirably, achieving accuracy rates of 98.8% and 98.4%, respectively. This research underscores the potential of machine learning algorithms in detecting and mitigating DDoS attacks. It emphasizes the necessity of employing advanced techniques for robust cyber threat defense and offers valuable insights into the performance of different machine learning algorithms in the context of DDoS attack detection.