A Study of Person Identification using Keystroke Dynamics and Statistical Analysis

In this paper, a basic study of closed-set identification using keystroke dynamics and simple statistical analysis has been carried out. Dwell time, flight time and one additional feature called key affinity are used as user-identifying features. The timing information is passed through a statistical layer to produce mean and standard deviation. This information is combined with key affinity to identify a rank-based person list. In conclusion, we compare the performance of this setup with other setups. This work aims to suggest that a keystroke dynamics system relying on pure statistics as its underlying algorithm may not be sufficiently accurate.