Real-time Visualization and Classification of DDoS Attack using Supervised Learning Algorithms

Abstract

With increased reliance on the Internet for various services, Distributed Denial of Service (DDoS) attacks are a major concern for organizations. DDoS attacks causes significant damage to the target system, leading to service downtime, data loss, and financial losses. To mitigate the impact of DDoS attacks, effective detection mechanisms are necessary. In this paper, a machine learning-based approach for DDoS attack detection is proposed. NSL-KDD Dataset of network traffic data containing both normal and attack traffic were used and various visualization techniques and machine learning models, including Random Forest Classifier, K-Nearest Neighbours Classifier and Logistic Regression were applied. The performance of these models was evaluated using crossvalidation and their accuracies were measured. The experimental results show that Random Forest Classifier outperforms other models with an accuracy of 99.33% before and after applying post pruning method. The experiment also focused on the false positives and false negatives and the implications of the results were discussed.