Exploring the Effectiveness of Machine and Deep Learning Techniques for Android Malware Detection

Abstract

The increasing occurrence of Android devices, coupled with their get entry to to touchy and personal information, has made them a high goal for malware developers. The open-supply nature of the Android platform has contributed to the developing vulnerability of malware assaults. presently, Android malware (AM) analysis strategies may be labeled into foremost categories: static evaluation and dynamic evaluation. These techniques are employed to analyze and understand the behavior of AM to mitigate its impact. This research explores the performance of DL model architectures, such as CNN-GRU, as well as traditional ML algorithms including SVM, Random Forest (RF), and decision tree (DT). The DT model achieves the highest accuracy (ACC) of 0.93, followed by RF (0.89), CNN-GRU (0.91), and SVM (0.90). These findings contribute valuable insights for the development of effective malware detection systems, emphasizing the suitability and effectiveness of the examined models in identifying AM.