Mobile Forensic Data Analysis: Suspicious Pattern Detection in Mobile Evidence

Abstract

Culprits' identi cation by the means of suspicious pattern detection techniques from mobile device data is one of the most important aims of the mobile forensic data analysis. When criminal activities are related to entirely automated procedures such as malware propagation, predicting the corresponding behavior is a rather achievable task. However, when human behavior is involved, such as in cases of traditional crimes, prediction and detection become more compelling. This paper introduces a combined criminal pro ling and suspicious pattern detection methodology for two criminal activities with moderate to the heavy involvement of mobile devices, cyberbullying and low-level drug dealing. Neural and Neurofuzzy techniques are applied on a hybrid original and simulated dataset. The respective performance results are measured and presented, the optimal technique is selected, and the scenarios are re-run on an actual dataset for additional testing and veri cation.