Mining mobile apps for anomalies

• Main Author: Seow, Wei Yang
• Other Authors: Shar Lwin Khin
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer systems organization::Special-purpose and application-based systems
• Online Access: http://hdl.handle.net/10356/76190
• Institution: Nanyang Technological University

There is no doubt that mobile applications play a huge role in our lives today, just in 2017 alone, there have been 178.1 billion mobile application downloads. [1] With our reliance on mobile applications, security of these applications is more important than ever, especially how applications handle your personal data. This project investigated the effectiveness of an approach consisting of a combination of topic modelling and static analysis to detect anomalous android mobile applications. We analysed applications from a repository known as fdroid and using Mallet, we split the applications into categories where we performed static analysis using FlowDroid on the applications. The source sink pairs of applications within each category were recorded down and a dataset was built. It was used to compare with unknown applications that belong in that category, via the source sink pairs to detect anomalies. Anomalous applications such as a repackaged application that contained malware was able to be detected. It was concluded that, the approach done in the project is viable, however there are still several aspects of the project that could be improved on to give better results.