Adaptive Static Analysis for Android Malware Detection: A Machine Learning-Based Security Framework

Users and organisations face serious security concerns as a result of the surge in malware threats brought on by the quick expansion of Android applications. Machine learning (ML)-based methods are a viable substitute for traditional signature-based malware detection systems, which find it difficult to keep up with changing threats. In this research, we investigate how to identify malware in Android apps using machine learning (ML) approaches. The study of various data pre-processing, dimensionality reduction, and classification methods is the main focus, and the generalisation capacity of the learnt models is evaluated using publicly available dataset. We place a strong emphasis on using feature selection (FS) techniques to minimise the dimensionality of the data and find the most pertinent features for classifying Android malware, which makes this process more explainable. Our methodology can determine which features are most essential to identifying an app as malicious. High accuracy in detecting malware in Android apps is achieved by the suggested method.