Unveiling Android Malware Behavior: A Hybrid Approach for Comprehensive Analysis and Classification
Sheikh Muhammad Zeeshan Javed, Muhammad Faisal Amjad · 2024
The increased usage of Android devices has increased data these devices manage in terms of volume and complexity, ranging from routine to highly sensitive information. The growing threats of Android malware require rigorous investigative approaches. Although significant progress has been made in understanding Android malware analysis, a considerable gap remains in behavioral dynamics. This research introduces a hybrid methodology through a detailed analysis of the behavioral patterns exhibited by various malware variants on Android devices. A comprehensive set of features, including sensitive permissions, network interactions, and API calls, are utilized in the categorization of malware. The VirusTotal API is employed to label malware samples, followed by deep-learning classifiers for malware classification. Statistical evaluation of the experimental results indicates that the proposed system achieves a notable prediction accuracy of 96%, demonstrating its effectiveness in malware classification.