Shared keystroke dataset for continuous authentication

Keystroke dynamics is an effective behavioral biometrics for user authentication at a computer terminal. Continuous or active authentication using keystroke dynamics has raised a lot of interest among researchers. However, there are only a few public datasets available for the research community compared to other biometric modalities primarily because of the difficulty of large scale data collection. Even the existing ones generally suffer from small number of subjects and lack of extensive features. In this paper, we provide the details on the collection of a shared dataset for the study of keystroke dynamics. We have collected raw keystroke data from 157 subjects allowing them to transcribe fixed text and answer questions freely. The dataset is characterized to reflect the temporal variations of typing patterns and the perturbations caused by different keyboard layouts. To show the usability and the quality of our dataset, we apply an existing algorithm, viz. Gaussian mixture model for keystroke analysis on the dataset and report the results.