Application to Identify the True Sender in Instant Messaging

Abstract

As cybercrime continues to spread, new e-forensic technologies are needed to counteract persistent yet anonymous fraud. The anonymity offered by the Internet has made the task of tracing criminal identity more difficult. Criminals use virtual identities to hide their true identities, for example via instant messaging (IM), which conceals inherent security risks and malicious tendencies from standard security software. It is essential that there are electronic forensic techniques to help identify cybercriminals as part of a criminal investigation, including identifying gender, writing habits, and stylometrics. This research paper uses statistical methods for the analysis of a template matching Pearson correlation coefficient or Euclid similarity to analyze Enron data and Viber chat messages with regard to distinct data in terms of the definition of the sender using neural networks. We also describe the structure of the creation and analysis of stylometric features for the cybercrime for electronic crime investigations. The proposed system has found a high rate of true identification of the real IM sender, using Pearson matching (88%) and an acceptable rate for Euclid similarity (72%).