Unifying Zero-Trust Architecture and Machine Learning for Continuous Threat Detection across Cloud-Based Enterprise Systems

Abstract

The increasing adoption of cloud-based enterprise systems has introduced complex cybersecurity challenges involving unauthorized access, identity compromise, insider threats, data breaches, and sophisticated network attacks. Traditional perimeter-based security approaches are insufficient for protecting distributed cloud environments where users, devices, applications, and workloads continuously interact across multiple networks. This research investigates the integration of Zero-Trust Architecture (ZTA) and machine learning (ML) to establish a continuous threat detection framework for cloud-based enterprise systems. The proposed approach combines identity verification, least-privilege access control, continuous authentication, behavioral analytics, and intelligent anomaly detection to strengthen enterprise cybersecurity. Machine learning algorithms analyze authentication records, network traffic, user activities, application logs, and access patterns to identify suspicious behavior and emerging threats. Zero-trust principles ensure that access decisions are continuously evaluated based on identity, device posture, contextual risk, and resource sensitivity. The research proposes an integrated security architecture that connects machine learning-based threat intelligence with adaptive access control and automated security response mechanisms. A quantitative experimental methodology is adopted to evaluate detection accuracy, false-positive rates, threat identification time, access decision efficiency, and system performance. The expected contribution is an adaptive, scalable, and intelligent cybersecurity framework that improves threat visibility, minimizes unauthorized access, and strengthens security resilience across enterprise cloud infrastructures.