Ransomware detection; machine learning; dataset

This dataset extracts the entropy of each of the PE sections of benign and ransomware reports to be used for detecting ransomware. Several machine learning classifiers were trained on this dataset such as Decision Tree, Random Forest, KNN, XGBoost and Naive Bayes. From the results, PE entropy can accurately detect ransomware with a decision tree classifier yielding the overall best result with a 98.8% accuracy and an AUC of 0.969. The latency with the prediction of the decision tree classifier was extremely quick with a result of 1.509 milliseconds.