MACHINE LEARNING AND DEEP LEARNING TECHNIQUES FORRANSOMWARE DETECTION: A COMPARATIVE STUDY

Abstract

Ransomware has emerged as one of the most destructive cybersecurity threats, necessitating intelligent and adaptive detection mechanisms beyond traditional security approaches. This study presents a comparative analysis of machine learning and deep learning techniques for ransomware detection, with emphasis on Artificial Neural Networks, Multilayer Perceptrons, Random Forest classifiers, and Long Short-Term Memory networks. The methodology adopted involved an extensive literature review of existing ransomware detection techniques, followed by an analytical comparison of traditional signature-based, behavior-based, and intelligent detection approaches. Experimental evidence from existing and proposed systems was examined to evaluate detection accuracy, precision, recall, and overall model performance. Feature-based datasets were analyzed, preprocessed, and evaluated using standard validation techniques to ensure consistency and reliability of results. The study highlights the strengths and limitations of each model, demonstrating that while machine learning algorithms offer efficiency and interpretability, deep learning models excel in capturing temporal ransomware behavior. Hybrid approaches combining deep learning and ensemble classifiers were found to deliver superior performance. The findings emphasize the necessity of intelligent ransomware detection frameworks capable of adapting to evolving attack patterns. This study contributes to cybersecurity research by providing a structured comparative framework and practical insights for selecting appropriate detection models