MACHINE LEARNING-DRIVEN TRUST SCORING FOR DYNAMIC FINANCIAL MICROSERVICE SECURITY

Abstract

Cloud-native financial institutions increasingly rely on distributed microservices to deliver scalable, resilient, and highly available banking, payment processing, insurance, and investment services. However, the dynamic nature of financial microservices introduces sophisticated cybersecurity challenges, including identity compromise, API attacks, insider threats, lateral movement, privilege escalation, and unauthorized service interactions. Conventional access control mechanisms primarily depend on static authentication and predefined security policies, limiting their ability to respond to continuously evolving cyber threats. Recent advances in Machine Learning (ML), Zero Trust Architecture (ZTA), behavioral analytics, cloud-native security, trust management, and predictive cybersecurity enable adaptive trust scoring for intelligent financial security. This paper proposes a Machine Learning-driven trust scoring framework for securing dynamic financial microservices by integrating continuous behavioral monitoring, adaptive trust evaluation, cloud-native monitoring, API security, and predictive threat analytics. The proposed framework continuously computes trust scores for users, workloads, APIs, devices, and service interactions before granting dynamic access permissions. Experimental analysis demonstrates significant improvements in cyberattack detection, adaptive access control, fraud prevention, operational resilience, regulatory compliance, and enterprise security compared with conventional rule-based security mechanisms. Keywords: Machine Learning, Trust Scoring, Financial Microservices, Zero Trust Architecture, Cloud Security, Behavioral Analytics, API Security, Identity Management, Financial Cybersecurity, Distributed Systems.