Leveraging Machine Learning Algorithms for Effective Android Ransomware Detection

The rising wave of Android ransomware attacks calls for sophisticated detection mechanisms to save its users from significant security threats. In this paper, the authors present an approach based on machine learning algorithms to identify Android ransomware, focusing on three: Random Forest, SVM, and Decision Tree. In this paper, three machine learning algorithms are trained with vast datasets of benign and malicious apps to learn patterns of behaviours that characterize ransomware. It was found that the Support Vector Machine was the best performer among them, showing the best accuracy, precision, recall, and F1 score; hence, it can minimize false positives and negatives effectively. Random Forest was close to SVM, followed by the Decision Tree, whose metrics were relatively lower. It calls attention to the potential of machine learning and, more specifically, SVM in remarkably enhancing the detection of Android ransomware, hence providing a more reliable and efficient security solution.