Behavioural-based malware detection on android phones

• Main Author: Kyran Ming Kuttan
• Other Authors: Liu Yang
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: https://hdl.handle.net/10356/148616
• Institution: Nanyang Technological University

The Android operating system is one of the most popular mobile operating systems in the market today. Applications developed using said operating system are continuously evolving and that include ones that have malicious intentions. There are many security measures put in place to prevent malware from being released into the application market, for instance permissions and Google Play Shield. However, malware continues to break through such methods as the development of malware continues to improve. In reaction, new methods of detecting malware have been researched to increase the effectiveness of malware detection. In this project, a methodology is proposed where the permissions used by an application is represented in the form of a graph, where the behaviour of an application can be seen. This form of graph can be termed as a permissions graph. An analysis is then conducted through the use of deep learning modes such as Feed-Forward Neural Network models and Neural Structured Learning (NSL) models. By using a permissions graph and an NSL model, the accuracy of detecting malware was desirable but can be improved on.