RANSOMWARE CLASSIFICATION USING MACHINE LEARNING

Abstract

In the context of increasingly common and sophisticated cyberattacks, ransomware has become a significant threat to information security, especially for organizations and businesses. Developing automated methods for detecting and classifying ransomware is crucial to minimizing financial losses and protecting reputations. This study focuses on classifying ransomware using machine learning models such as Random Forest, XGBoost, SVM, and Logistic Regression, leveraging three different datasets: the Android Ransomware Detection Dataset, BitcoinHeist Dataset, and Ransomware Detection Dataset. The research process includes data preprocessing, data balancing with SMOTE, feature selection, and hyperparameter optimization. The results indicate that Random Forest and XGBoost achieve the highest accuracy in classifying ransomware, with metrics such as accuracy, precision, recall, and F1-score outperforming SVM and Logistic Regression. This study provides a comprehensive view of the potential application of machine learning in ransomware detection while paving the way for future research and practical applications to enhance cybersecurity.