A Forensic Authorship Classification in SMS Messages: A Likelihood Ratio Based Approach Using N-gram

A Forensic Authorship Classification in SMS Messages: A Likelihood Ratio Based Approach Using N-gram Shunichi Ishihara The Australian National University School of Culture, History and Language Department of Linguistics, Canberra ACT 0200 Australia shunichi.ishihara@anu.edu.au Abstract Due to its convenience and low–cost, short message service (SMS) has been a very popu- lar medium for communication for quite some time. Unfortunately, however, SMS messages are sometimes used in illicit acts, such as com- munication between drug dealers and buyers, extortion, fraud, scam, hoax, false reports of terrorist threats, and many more. This study is a forensic study on the authorship classifica- tion of SMS messages in the Likelihood Ra- tion (LR) framework with the N–gram mod- elling technique. The aims of this study are to investigate 1) how accurately it is possible to classify the authors of SMS messages; 2) what degree of strength of evidence (LR) can be obtained from SMS messages and 3) how the classification performance and the LRs are affected by the sample size for modelling. The resultant LRs are calibrated by means of the logistic regress calibration technique. The re- sults of the classification tests will be rigor- ously assessed from different angles, using the techniques proposed for automatic speaker recognition and forensic voice comparison. 1 Introduction We often come across news stories on so–called cy- ber crimes which take advantage of the high visual anonymity of, for example, email and SMS mes- sages. In order to combat these cyber crimes, the Australian Government is currently trying to pass the Cybercrime Legislation Amendment Bill 2011 (hereafter, Cyber Law). This Cyber Law was in- troduced and read for the first time at the House of Representatives in June, 2011.1 This legislation will enable police and intelligence agencies to in- struct phone companies and internet carriers not to destroy sensitive information, such as text messages or emails from terrorists or criminals, that is impor- tant for investigations and prosecutions. This leg- islation also set the framework for Australia to join the Council of Europe Convention on Cybercrime, which more than 40 nations have joined or plan to join. Needless to say, SMS messages, which are the fo- cus point of the current study, hold a very impor- tant position in the above–mentioned legislation. As Grant (2007, p2) states “[o]ver recent years there has been considerable and growing interest in forensic authorship analysis”, it is predicted that SMS mes- sages will be increasingly used as evidence in Aus- tralian courts and in national and international secu- rity contexts (Coulthard and Johnson, 2007).2 The fact that the use of mobile phones has been increas- ing exponentially and that the SMS is becoming a more and more common medium of communication, is apparently a strong driving force and motivation for the above–mentioned legislation and the conduct of fundamental research on SMS messages as scien- tific evidence. Having said that, there is a large amount of re- search on authorship attribution in general (Thisted and Efron, 1987; Pennebaker and King, 1999; Dod- dington, …

Content truncated to 500 words