Longitudinal Keystroke Biometric Studies on Long-Text Input
Elizabeth Anne Wood, Julio Zelaya, Eric Saari, Kenneth King, Mike Gupta, Nicola Howard, Sadia Ismat, Mary Angela Kane, Mark Naumowicz, Daniel Varela, Mary Villani · 2008
Pace University’s keystroke biometric system is able to identify individuals with a high degree of accuracy. A relatively new dichotomy model system has been developed to support studies of keystroke biometric authentication, and preliminary results with this system have shown promise. The current study extends those authentication experiments and examines the consistency of an individual’s keystroke patterns over periods of week and years, expanding upon a small preliminary longitudinal study carried out on four subjects in the fall of 2007. Results of the preliminary study suggest that an individual’s typing patterns do not change significantly over a period of 2 to 4 weeks. The current study repeats the fall 2007 experiments with a larger sample size, and begins to examine the consistency of typing patterns over a longer period of approximately two years. The study also provides new data on keystroke patterns in a set of identical twins, shedding light on the extent to which keystroke patterns may be biologically determined. 1.