Two Methods for Computer Users Identification

摘要

We present the use of back propagation neural network and correlation analysis in order to verify the identity of computer users. User data are monitored at the keyboard level, by performing real-time measurements of the lime duration's between the characters entered in user text (keystroke dynamics). All monitored users entered the same fixed phrase in our experiments. Part of the collected data was used to train the decision procedures, the rest of data served for testing. Achieved results confirm suitability of presented approaches in a secondary authentication procedure.