Ransomware Detection using ML

Abstract

This study investigates the effectiveness of three machine learning classifi-ers—Random Forest, SVM, and XGBoost—in detecting benign files using a dataset of file features. The dataset is cleaned by removing rows with missing values and du-plicates, followed by feature scaling with StandardScaler. Correlation heatmaps and confusion matrix visualizations are used to explore data relationships and model performance. The classifiers are trained and evaluated based on accuracy, precision, recall, F1-score, and ROC-AUC. Results indicate the comparative performance of each model, highlighting their strengths and weaknesses in distinguishing between benign and non-benign files. This comprehensive approach provides insights into the most suitable classifier for this specific detection task.