AI-driven intrusion detection system for cloud security

This Intrusion Detection Systems (IDS) are essential for safeguarding modern cloud infrastructures, yet traditional signature-based methods struggle to detect evolving and zero-day cyber threats. Building upon the AI-driven intrusion detection concepts outlined in the base paper

, this work presents an enhanced, multi-layered IDS framework integrating both Machine Learning and Deep Learning techniques for robust threat identification. The proposed system employs Random Forest (100-tree) and Support Vector Machine with RBF kernel for high-precision classification, while LSTM networks capture sequential attack patterns and autoencoders perform anomaly detection on complex traffic distributions. A hybrid weighted-voting ensemble further improves detection reliability by combining heterogeneous model outputs. The system is implemented for real-time intrusion monitoring, providing continuous analysis of streaming network traffic and instant alert generation through a live visualization dashboard. Experimental results demonstrate significant improvements in accuracy, adaptability, and false-positive reduction compared to standalone models and traditional IDS approaches. This integrated AI-based IDS architecture offers a scalable and intelligent solution for securing cloud environments against advanced cyber threats.

Keywords:
Artificial Intelligence, Intrusion Detection System (IDS), Cloud Security, Machine Learning, Random Forest, Support Vector Machine (SVM-RBF), Deep Learning, LSTM Network, Autoencoder, Anomaly Detection, Hybrid Ensemble, Real-Time Monitoring, Cybersecurity, Network Traffic Analysis