An Effective Classification of DDoS Attacks in a Distributed Network by Adopting Hierarchical Machine Learning and Hyperparameters Optimization Techniques

Abstract

Data privacy is essential in the financial sector to safeguard sensitive client information, prevent financial crimes, ensure regulatory compliance, and protect intellectual property. The rise of digital transactions and internet usage has made securing client data more challenging, with Distributed Denial of Service (DDoS) attacks emerging as a significant threat to privacy. To address this, efficient and resilient detection and prevention systems are critical. Machine learning offers a promising solution for developing effective cyberattack detection models, enhancing the ability to protect sensitive information and contribute to both scientific and practical advancements. The CIC-IDS 2017 dataset, known for its comprehensive collection of network traffic data, was used for model training and evaluation. To improve model performance, feature selection was performed using Lasso, while oversampling was applied through SMOTE to address data imbalance. Various machine learning algorithms were employed, with and without hyperparameter optimization. Among the models tested, the best performance was achieved by a voting classifier, which attained 100% accuracy, demonstrating the robustness and effectiveness of ensemble methods in detecting DDoS attacks.