Using Alert Levels to enhance Keystroke Dynamic Authentication
Alex Andersen, Simen Hagen · 2007
Authentication and verification of users in computer security are areas which gains a lot of attention. A reason for this is the high number of inside attacks, where already authenticated user accounts are used to gain access to prohibited information or privileges. Session hijacking, password stealing/guessing or perimeter possession are examples of areas where ordinary authentication has been known to fail. A secret password and public username is the most widespread authentication and verification scheme used. This research will purpose to add layers of software biometrics into the authentication and verification process to increase security. This will be done using keystroke dynamics, which is a way to distinguish a users pattern of typing, giving or removing privileges. Making use of Alert levels, can decrease the Mean Error Rate and especially the False Rejection Rate, when accepting some human behavior anomalies. 1