Comprehensive Android Malware Detection: Leveraging Machine Learning and Sandboxing Techniques through Static and Dynamic Analysis

Android malware has grown alongside Android smartphones in recent years. One of the fastest-growing sectors worldwide is the Internet of Things (IoT). The large number of third-party apps that benefit users makes Android the most popular smartphone operating system worldwide. This research investigates using machine learning algorithms to improve Android malware detection. Although successful, traditional signature-based detection technologies cannot keep up with the rapid growth of malware. In contrast, machine learning can identify malware patterns and behaviors, providing a more dynamic and robust solution. Using supervised learning models including k-nearest neighbour(kNN), Random Forests (RF), Decision tree (DT), and Naive Bayes (NB), this study examines machine learning methodologies. In this proposed work, we analyze machine learning methods using static and dynamic properties extracted by Androguard and Droidbot. Outperforming other classification models, the kNN classifier has 99% accuracy. Decision tree may reach 98% accuracy, whereas Random Forest could reach 92%. But the Naive Bayes classifier had an 86% accuracy.