Supervised Learning Approaches in Android Malware Detection: Survey and Analysis

The exponential growth of Android-based mobile applications has led to a parallel surge in sophisticated malware, necessitating more adaptive and intelligent detection strategies. While numerous studies have explored supervised machine learning (ML) for Android malware detection, most existing reviews either generalize across learning paradigms or focus narrowly on specific detection phases. This paper presents a comprehensive and focused survey of Android malware detection techniques based solely on supervised learning approaches. Unlike previous studies, it introduces a refined classification framework that distinguishes detection strategies based on analysis types, feature engineering depth, and learning model adaptability. Additionally, this work consolidates and contrasts findings from a wide range of recent literature, providing deeper insight into architectural trends, performance benchmarks, and practical deployment challenges. By analyzing the trade-offs between accuracy, interpretability, and computational cost across static, dynamic, and hybrid detection paradigms, this study contributes a well-structured reference point for researchers aiming to build scalable and robust malware defense systems.