On Static Android Malware Detection and Analysis: A Systematic Review

Since the past few years, the number of smartphone users have significantly risen, specifically the Android users owing to being economically feasible making up a huge market share. Owing to its architecture and popularity, the Android platform is disposed to malware attacks. The detection of malware is becoming the primary concern on Android platforms. Machine learning (ML) and deep learning (DL)-based algorithms have been able to successfully detect malwares on Android with high degree of accuracy, outperforming other existing traditional methods. This paper proposes a taxonomy of various Android malware detection and analysis techniques. The detection techniques based on signatures, heuristics, behaviour, machine learning, and anomaly detection have been discussed in this paper. It presents an extensive literature survey from reliable databases that includes Springer, Web of Science, Science Direct, IEEEXplore, and Google Scholar for the systematic analysis. This paper also highlights the open issues present within existing literature and illustrates an outline for the proposed work that can further enhance automated procedures in the detection of static malware on Android. Further, various Android malware analysis schemes like static, dynamic, and hybrid have also been discussed. Based on learning algorithms such as Linear Regression, Random Forest, Deep Neural Network (DNN), Support Vector Machine (SVM) and Naïve Bayes (NB), etc. A number of researchers have worked on different algorithms on different features of Android static malware to detect malware, the review of which has been presented systematically in this study.