Collecting and Leveraging Identity Cues using Keystroke Analysis (CLICKA)
Keystroke dynamics is the analysis of the way in which an individual types. This can be used to draw identifying characteristics about an anonymous end-user. The typing behaviours displayed by an individual can be as uniquely identifiable as their handwriting or their signature. There has been significant research into analysing the typing rhythm and cadence to identify a specific user. Similarly, these techniques have also been exploited to derive physical attributes and demographic information about the user at the keyboard, such as their hand size, handedness and typing style. The research in this proposal is significantly different from previous research which merely attempted to identify the characteristics of individual users, i.e. to prove that a particularly user is using the keyboard. Our research attempts to learn real-world identity ‘cues’ about the user and hence learn characteristics such as name, native language etc.
This work investigated the repeatable and predictable typing behaviours based on their familiarity with the data. In essence this means that it is possible to discern the words or strings of characters that are typed more frequently by an individual, for example, a user’s name.
CLICKA builds upon these ideas to determine identity ‘cues’ about an anonymous user, based solely on analysis of their typing patterns. In the first instance the research developed an experimental method to determine the name of an anonymous user. The research then evolved this idea a step further to determine the native language of an individual, using a similar underlying idea; there are elements and combinations that uniquely popular in certain languages and as such it is expected they would be identifiable, for example ‘er’ is very common in the English language.