AI AND MACHINE LEARNING-DRIVEN FRAMEWORK FOR EARLY DETECTION AND PREVENTION OF RANSOMWARE ATTACKS IN BANKING SYSTEMS

Abstract

Ransomware has emerged as one of the most persistent and destructive forms of cyber threats, targeting critical infrastructures, financial institutions, and data-dependent enterprises. Traditional detection mechanisms are often reactive and fail to respond effectively to new and evolving ransomware variants. This study presents an AI and Machine Learning (ML)-driven framework for the early detection and prevention of ransomware attacks, focusing on its deployment within the banking sector. The proposed framework integrates supervised learning models with feature extraction and behavioral analysis to identify anomalous activity indicative of ransomware. Experimental evaluation using a simulated dataset demonstrates that the proposed model achieves a detection accuracy of 96.4%, with an area under the ROC curve (AUC) of 0.94, outperforming conventional signature-based and heuristic techniques. The results confirm that AI-driven behavioral detection provides superior adaptability and precision against zero-day ransomware variants. This research contributes to the growing body of cybersecurity studies by offering an empirically validated framework capable of real-time threat identification and risk mitigation in financial systems.