CAGDEEP : Mobile malware analysis using force atlas 2 with strong gravity call graph and deep learning

• Main Authors: Nur Khairani, Kamarudin, Ahmad Firdaus, Zainal Abidin, Azlee, Zabidi, Mohd Faizal, Ab Razak
• Format: Conference or Workshop Item
• Language: English; English
• Published: Institute of Electrical and Electronics Engineers Inc. 2023
• Subjects: QA75 Electronic computers. Computer science; QA76 Computer software; T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/40367/1/CAGDEEP_Mobile%20malware%20analysis%20using%20force%20atlas%202.pdf; http://umpir.ump.edu.my/id/eprint/40367/2/CAGDEEP_Mobile%20malware%20analysis%20using%20force%20atlas%202%20with%20strong%20gravity%20call%20graph%20and%20deep%20learning_ABS.pdf; http://umpir.ump.edu.my/id/eprint/40367/; https://doi.org/10.1109/ICSECS58457.2023.10256350

Today many smart devices are running on Android systems. With the increasing popularity of Android, mobile malware continuously evolves as well, and further attacks Android operating systems. To address these shortcoming issues many security experts use different approaches to detect malware based on various static features. However, by considering only the statistical features, the potential semantic information such as the behavioral feature of the code is overlooked. To leverage the existing static analysis techniques, this study proposes CAGDeep, to reflect deep semantic information of malware samples. The novelty of our study lies in the Force Atlas 2 call graph development to capture malware behavior patterns. Afterwards, this study adopts Convolutional Neural Network (CNN) for malware detection and classification algorithm. We compare CAGDeep with a state-of-the-art Android malware detection approach. Our evaluation results demonstrate that CAGDeep can achieve 80% accuracy for detecting malware.