Abstract:
As the usage of Android mobile devices has increased rapidly, so has the occurrence of Android malware. To mitigate the risk posed by Android malware, it is important to develop effective de- tection techniques. In this paper, we propose a machine learning and deep learning-based approach for Android malware detection that involves static analysis of malware. We also have used a wide variety of learning-based methods to build detection models using the dataset and further leverage explainability algorithms such as LIME and SHAP. We provide a methodology to utilize the core functionality of explainability algorithms in making lightweight, efficient, and temporally stable models, besides increasing the users’ trust. Our experimental results demonstrate the effectiveness of our approach in accurately detecting both benignware and malware. Furthermore, among malware samples, we classified the samples by Malware Family and Category. The proposed approach can be useful for security practitioners, researchers, and developers in building more secure and resilient Android applications.
