Optimizing Neural Network Architecture for Detecting DDOS Attacks using ANN and XGBoost in Imbalanced Networks

This research addresses the critical challenge of detecting Distributed Denial of Service (DDoS) attacks, specifically targeting HTTP-based services within highly imbalanced network environments. Traditional detection mechanisms often struggle with high false-alarm rates and poor sensitivity when benign traffic heavily outweighs attack signatures. To resolve this, the paper presents a hybrid machine learning framework that integrates Artificial Neural Networks (ANN) with XGBoost, optimized via Particle Swarm Optimization (PSO) and balanced using a DBSCAN-SMOTE preprocessing pipeline. In the proposed methodology, DBSCAN-SMOTE addresses class imbalance by synthesizing minority class samples while filtering out noise. The customized ANN architecture acts as the primary classifier to distinguish normal traffic from malicious patterns, while the subsequent

XGBoost layer refines these predictions to maximize classification accuracy. According to the paper's reported findings, this integrated DBSCAN-SMOTE-ANN-XGBoost-PSO model achieves a high accuracy of 96.83%, a sensitivity of 93.23%, and a precision of 96.13%. For M.Tech researchers, this study provides a robust blueprint for developing hybrid intrusion detection systems, offering valuable guidance on hyperparameter tuning, feature selection, and handling skewed class distributions in real-world network datasets.

Message: Hi FE, I need help with “Optimizing Neural Network Architecture for Detecting DDOS Attacks using ANN and XGBoost in Imbalanced Networks” in “Computer Science & Engineering”