Year: 2026
Type: article
Abstract: Ransomware is a type of malicious software that encrypts important user files and demands a ransom to restore access, posing a serious threat to individuals, businesses, and organizations.Traditional antivirus systems primarily rely on signature-based detection techniques, which may fail to identify new and unknown ransomware variants.To address this challenge, this paper proposes a Machine Learning-Based Framework for Ransomware Detection and Classification that uses behavior-based analysis to identify potentially malicious files.The proposed framework involves dataset collection, data preprocessing, feature extraction, and machine learning model training.Relevant features, including file size, file entropy, API calls, file operations, CPU usage, memory usage, and network activity, are analyzed to identify patterns associated with ransomware behavior.A Random Forest classifier is used to classify input files or their extracted features as benign or potentially malicious.The trained model is integrated with a Flask-based web application that allows users to submit files for analysis and view the detection results.The framework also categorizes detected threats into high, medium, and low risk levels to help users understand potential security risks.The proposed approach aims to improve the identification of suspicious files, support early threat detection, and reduce the risk of data loss.This framework demonstrates the potential of machine learning in strengthening cybersecurity and provides a foundation for future enhancements, including real-time monitoring, improved classification models, and automated threat response mechanisms.
Source: IJARCCE
Authors: Gnanamani H, Akshata G B, Ananya K, Anju K, Anusha A
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7235198383
