SMS linguistic profiling authentication on mobile device

It is commonly acknowledged that mobile devices now form an integral part of an individual's everyday life. As the amount of valuable and sensitive information stored on a mobile device increases, so does the need for effective security. In order to protect unauthorised access, an authentication system is required. Biometric authentication has proven to be a more reliable solution than knowledge based and token based techniques. Indeed, biometric techniques are uniquely individual, impossible to forget or lose, difficult to reproduce or falsify and difficult to change or hide. Despite the well known advantages of biometric authentication approaches, the majority of current state-of-the-art mobile devices embrace point of entry authentication systems including PIN/passwords and one time fingerprint verification. This paper introduces a feasibility study into a novel biometric technique for mobile devices, linguistic profiling. This investigation sought to authenticate users based on their writing vocabulary and style of SMS messages. The findings, based upon 30 participants, revealed the feasibility of the approach. While an overall average Equal Error Rate (EER) of 24% is unacceptably high, several users experienced an EER of 0%, suggesting significant potential to apply the technique for a subset of the population.