Enhancing Network Intrusion Detection System Using NDAE With Xgboost

This study addresses the challenge of enhancing Network Intrusion Detection Systems (NIDS) by integrating Non-symmetric Deep Auto-Encoders (NDAE) with eXtreme Gradient Boosting (XGBoost). The study aimed to develop a system capable of efficiently processing large volumes of data for more accurate threat detection. Methodologically, it employed the KDD Cup 99 benchmark dataset, with preprocessing to correct class imbalances, and combined the feature representation strength of NDAE with the classification accuracy of XGBoost. The findings suggest that while the model underperforms in specific metrics, it significantly reduces data processing time and improves imbalanced data classification. The study concludes that despite limitations like class imbalance, the hybrid model shows promise for rapid threat detection. This research contributes to the field by demonstrating a scalable and efficient approach to NIDS, with future work suggested in advanced machine learning techniques for improved predictive accuracy.