An Intelligent Machine Learning Framework for Adaptive Cyber Threat Detection Across Multi-Cloud Computing Environments

Abstract

The rapid expansion of cloud computing has encouraged enterprises to adopt multi-cloud architectures to improve service availability, workload distribution, scalability, and business continuity. Despite these advantages, managing security across multiple cloud providers remains a significant challenge because each platform generates diverse security logs, network events, and access records. The fragmented nature of these environments makes it difficult for conventional security solutions to correlate security incidents and detect sophisticated cyberattacks in a timely manner. Consequently, there is a growing need for intelligent security frameworks capable of providing unified threat analysis across heterogeneous cloud infrastructures.This study proposes an adaptive artificial intelligence-based threat detection framework that enhances cybersecurity in multi-cloud environments through automated data analysis and intelligent decision-making. The proposed framework aggregates security-related information from multiple cloud resources, preprocesses the collected data, extracts discriminative features, and applies supervised machine learning techniques to classify benign and malicious activities. By learning behavioral characteristics from historical security events, the system effectively identifies anomalous network traffic, unauthorized authentication attempts, suspicious user behavior, privilege misuse, and other potential security threats. Unlike conventional rule-based approaches, the proposed framework continuously improves its detection capability by recognizing evolving attack patterns and reducing dependence on manually defined security rules.The effectiveness of the proposed framework is evaluated using standard machine learning performance metrics, including accuracy, precision, recall, F1- score, and false-positive rate. Experimental analysis demonstrates that intelligent learning models can significantly improve threat detection capability while minimizing false alarms and maintaining consistent performance in dynamic cloud environments. The proposed framework offers a scalable, efficient, and proactive security solution that strengthens cyber resilience, enables early threat identification, and supports secure service delivery across distributed multi-cloud computing platforms.