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

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.