DDoS Attack Classification Using Machine Learning

Abstract – DDoS (Distributed Denial of Service) attacks pose a significant threat to network security by overwhelming target systems with excessive traffic, rendering them inaccessible to legitimate users. Traditional rule-based detection methods struggle to keep up with evolving attack patterns, making machine learning (ML) a promising alternative. This study explores various ML models, including Random Forest, K- Nearest Neighbors (KNN), XGBoost, and Decision Tree classifiers, to detect and classify DDoS attacks effectively. Using a dataset containing over 852,585 network traffic records, we preprocess the data through feature selection, encoding, and outlier removal before training and evaluating different models. Performance is assessed using metrics such as accuracy, F1-score, recall, and confusion matrices. The results indicate that the Random Forest classifier achieves the highest accuracy (99.82%), followed closely by Decision Tree and KNN. The study highlights the effectiveness of ML in DDoS detection and suggests future improvements, such as hyperparameter tuning, deep learning techniques, and real-time deployment for enhanced security. Index Terms – Random Forest, KNN, SVM, Decision Tree.