De-Anonymization of Anonymous E-mails in digital forensic

Abstract

Rapid growth of e-mail users and popularity of Internet also increase the number of cybercrime related with e-mail. Anonymity provides sophisticated mechanism to preserve privacy of user’s online life. Internet has become an ideal place for criminals over the past decades due some of its characteristics. Misuse of Anonymity by Cyber criminals for their wrong doings is a threat to regular online users. E-mail, the modern solution to the snail mail is the most frequent method of cybercrime communication. Individual writing style may contain hidden patterns that can use to uniquely identify an individual using his writing samples. Traditional method of digital forensic on e-mail does not address the domain of authorship analysis. E-mail header only based e-mail forensics may mislead the entire investigation towards a wrong direction and may prosecute the innocent at the end. Mining of information from the e-mail text will provides the knowledge about the true authorship of the text. Mining of information on e-mail can use statistical or machine learning based methods. Use of content based authorship analysis in e-mail forensic will increase the probability of proving guiltiness of a criminal in the court of law. This project used computational stylometry and related text mining techniques in natural language processing to analyze and identification of authors writing discriminators for recognize the best matching author of an anonymous e-mail. Association rule mining algorithms such as Apriori can use to identify the frequent writing features from personal writing samples and it reduce the burden of stylometry feature analyses. Experimental results conclude the effectiveness of the proposed mechanism over traditional approaches. Result of this project proves that there are stylometry patterns that present uniquely in person’s writing and they can use to identify the cyber criminals who uses email in their activities.