ZeTTA-IDS: Zero-Trust Threat-Aware Intrusion Detection System Using Cyber Threat Intelligence

Machine Learning–based Intrusion Detection Systems (IDS) are effective in identifying network attacks, but their performance is often limited by static training models and the assumption that Cyber Threat Intelligence (CTI) sources are inherently trustworthy. Existing CTI-integrated IDS frameworks typically rely on single-source intelligence and lack mechanisms to evaluate the reliability and consistency of CTI data, making them vulnerable to false positives and data poisoning attacks. This study proposes a Trust-Aware CTI-Integrated Adaptive IDS that enhances detection performance through reliable CTI fusion, explainable decision-making, and continuous online learning. The system introduces a trust and reliability scoring mechanism that evaluates CTI based on factors such as source reputation, recency, consensus, and behavioral consistency with observed network traffic. A multi-source CTI fusion model combines intelligence from various feeds and resolves conflicts using trust-based weighting. Additionally, a CTI Transfer Model converts validated intelligence into high-quality training data for secure and adaptive model updates. The integration of explainable AI improves transparency, while Zero Trust principles enable adaptive, risk-based responses. Experimental results demonstrate that the proposed model achieves 96.6% accuracy and a reduced false positive rate of 4.4%, significantly outperforming traditional and standalone ML-based IDS approaches.