Startseite // SnT // Research // SerVal // Projects // Hinting at Malicious Code in Android Apps with Graph and Data Clustering Techniques

HitDroid: Hinting at Malicious Code in Android Apps with Graph and Data Clustering Techniques

Project title: HitDroid: Hinting at Malicious Code in Android Apps with Graph and Data Clustering Techniques
Principal investigator: Jacques Klein
Vice principal investigator: Tégawendé Bissyandé
Funding: University of Luxembourg Internal Project
Starting date and Duration: 01/07/2018 for 3 years
Contact personJacques Klein

In HitDroid, we aim at proposing a new malware detection approach in mobile ecosystems. It builds on our previous work investigating millions Android apps to reveal how malware is mostly built via repackaging popular apps. Although the community is slowly focusing on repackaged app detection, the existing approaches remain impractical because they require pairwise comparison of apps.

We propose an original detection approach that analyses each app alone, making the approach practical, leveraging graph mining techniques to identify alien code grafted to an “original” code graph.