User relationship classification of facebook messenger mobile data using WEKA

Publication Type:
Conference Proceeding
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2018, 11058 LNCS pp. 337 - 348
Issue Date:
Full metadata record
Files in This Item:
Filename Description Size
user-relationship-classification-cameraready (3).pdfAccepted Manuscript version1.48 MB
Adobe PDF
© Springer Nature Switzerland AG 2018. Mobile devices are a wealth of information about its user and their digital and physical activities (e.g. online browsing and physical location). Therefore, in any crime investigation artifacts obtained from a mobile device can be extremely crucial. However, the variety of mobile platforms, applications (apps) and the significant size of data compound existing challenges in forensic investigations. In this paper, we explore the potential of machine learning in mobile forensics, and specifically in the context of Facebook messenger artifact acquisition and analysis. Using Quick and Choo (2017)’s Digital Forensic Intelligence Analysis Cycle (DFIAC) as the guiding framework, we demonstrate how one can acquire Facebook messenger app artifacts from an Android device and an iOS device (the latter is, using existing forensic tools. Based on the acquired evidence, we create 199 data-instances to train WEKA classifiers (i.e. ZeroR, J48 and Random tree) with the aim of classifying the device owner’s contacts and determine their mutual relationship strength.
Please use this identifier to cite or link to this item: