Use of Correlation among Keystroke Biometric Features for Handling Missing or Infrequent Data

Steve H Kim, Pace University


As more individuals and companies rely on electronic resources to maintain sensitive information, the security of those resources becomes exponentially more important. Despite the growing importance of network security cost prohibitions and underdeveloped technology have left users and corporations alike relying on outdated unsecure methodologies to protect electronic resource outlets. Properly sampled and analyzed keystroke biometrics presents a viable and cost effective system of user authentication, with the benefit of exponentially increased security. In earlier keystroke biometric feature extraction, two types of fallback models were studied to handle missing or insufficient key input data. One was the linguistic model and the other was the keyboard physiological model. Here both models are criticized by studying the poor linear correlation between a missing key and its parent. Here, a correlation based fallback table model based on the linear correlation coefficients among different keys is proposed as a new statistical fallback model. A large long text keystroke database is used to construct the model and a short free text keystroke database is used to reveal the superiority of the proposed model over the two previous models.

Subject Area

Computer science

Recommended Citation

Kim, Steve H, "Use of Correlation among Keystroke Biometric Features for Handling Missing or Infrequent Data" (2013). ETD Collection for Pace University. AAI22623093.



Remote User: Click Here to Login (must have Pace University remote login ID and password. Once logged in, click on the View More link above)