A Model for User Authentication Based on Manner of Keystroke and Principal Component Analysis

As information and network security is exposing to increasing evil threats significantly, authentication is now playing a more and more important role in security defense system. In this paper, a model that authenticates and identifies access to computer systems and networks by applying keystroke manner is put forward. In the model, legitimate user´s typing manner such as durations of keystrokes and latencies between keystrokes are collected during training step, and several features are extracted from the keystroke data. Then user´s manner template of keystroke is built by principal component analysis. While identifying, some principal component scores of user data are used to judge the user´s validity. The procedure of authentication is invisible to user. The results of related experiment show that the model has a good discerning ability