Résumé
E-mail has become the most common way to communicate on the Internet, but e-mail security and privacy mechanisms are still lacking. This has proven to be a very valuable characteristic for criminals, who can easily take advantage of e-mail’s various weaknesses to remain anonymous. Consequently, cybercrime investigators need to rely on computer-aided writeprint modelling methods and tools to identify the real author of malicious emails with transformed semantic content. In this paper, we propose a customized version of associative classification, a well-known data mining method, as well as a Support Count method, to address the authorship attribution problem. Experimental results on real-life data suggest that our proposed algorithms can achieve good classification accuracy on the e-mail author attribution problem through the use of writeprint modeling.