Exposing Android Ransomware using Machine Learning

摘要: The Ransomware detection reports from cyber-security companies trigger high threat in Android devices vulnerability. study used machine learning approaches, particularly classifiers: Decision Tree, Random Forest, Gradient Boosting Trees, and AdaBoost to detect malware. dataset HelDroid with known Ransomware's features, the was transformed feed on classifier model. Using 5-attribute classifier, models generate average of 98.05% accuracy rate, both training test sets. same results Naive Bayes classifiers mean cross-validation Gaussian Bernoulli is 97.6%, while Multinomial 81.6%. Feeding binarized 229-attribute dataset, Tree generates 99.08% accuracy, three Classifiers returns 100% overfit results.