An Experimental Study on Authorship Identification for Cyber Forensics

Abstract

Authorship Identification is subfield of authorship analysis deals with finding the plausible author of anonymous messages. The Authorship identification problem of online messages is challenging task because cyber predators make use of obscurity of Cyberspace and conceal the identity. By performing the forensic analysis of online messages, empirical evidence can be collected. These evidences can be used to prosecute the cybercriminal in a court and punish the guilty. This way cybercrimes can be minimized up to certain extent by detecting the true indentities.Therefore it is required to build up innovative tools & techniques to appropriately analyze large volumes of suspicious online messages. This paper compares the Performance of various classifiers in terms of accuracy for authorship identification task of online messages. Support Vector Machines, KNN, and Naïve Bayes classifiers are used for performing experimentation .This paper also investigate the appropriate classifier for solving authorship of anonymous online messages in the context of cyber forensics.