AI-Based Intrusion Detection Systems Using Ensemble Deep Learning Models

This study investigates the development and evaluation of AI-based intrusion detection systems (IDS) leveraging ensemble deep learning models. In an era of increasing cyber threats, traditional IDS often struggle with high false positive rates and inadequate adaptability to evolving threats. The proposed research integrates Convolutional Neural Networks (CNN), Long Short-Term Memory networks (LSTM), and Autoencoders within an ensemble framework to enhance anomaly detection in network traffic. Using the NSL-KDD dataset, performance metrics such as accuracy, precision, recall, F1-score, and Area Under the Curve (AUC) were analyzed. The ensemble model demonstrated superior performance compared to individual deep learning models. Statistical techniques including regression analysis and ROC curve evaluation confirmed the ensemble model's robustness. This approach provides a scalable and adaptive solution for securing networks against sophisticated cyberattacks. Keywords Intrusion Detection System, Ensemble Deep Learning, Cybersecurity, CNN, LSTM, Autoencoder, Anomaly Detection