In order to gain information about the context, purpose, and functionality of mobile applications and to classify them accordingly, this study examines a variety of application metadata.
In mobile application analysis, the focus is usually placed on an approximated version of the original program code. The collection of safety-relevant properties usually succeeds only individually per application and does not allow a broader view of which properties are characteristic for the functionality of applications. In this project application analysis was approached from a new perspective in order to draw conclusions about applications based on their metadata.
The goal of the project was to aggregate application-related information, off-program, such as GUI layouts or application text, and then aggregate it to create a basis of distinction for individual apps based on that metadata. The approach designed in this way can help identify individual types of applications and find semantically similar ones. In particular with regard to malware or applications with security relevance, this results in a novel approach to understanding their relation to other applications and classifying them accordingly.