A Comprehensive Review of Android Security: Threats, Vulnerabilities, Malware Detection, and Analysis. (29th June 2022)

• Record Type: Journal Article
• Authors: Acharya, Saket; Rawat, Umashankar; Bhatnagar, Roheet
• Other Names: Bhushan Bharat Academic Editor.
• Is Part Of: Security and communication networks. Volume 2022(2022)
• Publication Date: 2022-06-29
• Subjects: Computer networks -- Security measures -- Periodicals; Computer security -- Periodicals; Cryptography -- Periodicals; 005.805
• DOI: 10.1155/2022/7775917
• Languages: English
• ISSNs: 1939-0114

Abstract

Abstract : The popularity and open-source nature of Android devices have resulted in a dramatic growth of Android malware. Malware developers are also able to evade the detection methods, reducing the efficiency of malware detection techniques. It is hence desirable that security researchers and experts come up with novel and more efficient methods to analyze existing and zero-day Android malware. Most of the researchers have focused on Android system security. However, to examine Android security, with a specific focus on malware development, investigation of malware prevention techniques and already known malware detection techniques needs a broad inclusion. To overcome the research gaps, this paper provides a broad review of current Android security concerns, security implementation enhancements, significant malware detected during 2017–2021, and stealth procedures used by the malware developers along with the current Android malware detection techniques. A comparative analysis is presented between this article and similar recent survey articles to fill the existing research gaps. In the end, a three-phase model is proposed to efficiently identify and characterize Android malware. In the first phase, a lightweight deep transfer learning approach is used to classify Android applications into benign and malicious. In the second phase, the malicious applications are executed in a virtual emulator to reduce the number of false positives. Finally, the malicious applications having the same characteristic ratio are grouped into their corresponding families using the topic modelling approach. The proposed model can efficiently detect, characterize, and provide a familial classification of Android malware with a good accuracy rate.