User identification through usage analysis of electronic devices

Abstract

Different aspects of usage of electronic devices significantly vary person to person, and therefore, rigorous usage analysis exhibits its prospect in identifying a user in road to secure the devices. Different state-of-the-art approaches have investigated different aspects of the usage, such as typing speed and dwelling time, in isolation for identifying a user. However, investigation of multiple aspects of the usage in combination is yet to be focused in the literature. Therefore, this paper, we investigate multiple aspects of usage in combination to identify a user. We perform the investigation over real users through letting them interact with an Android application, which we develop specifically for the investigation. Our investigation reveals a key finding considering multiple aspects of usage in combination provides improved performance in identifying a user. We get this improved performance up to a certain number of aspects of usage being considered in the identification task.