Evaluating the Reliability of Credential Hardening through Keystroke Dynamics

Abstract

Most computer systems rely on usernames and passwords as a mechanism for authentication and access control. These credential sets offer weak protection to a broad scope of applications with differing levels of sensitivity. Traditional physiological biometric systems such as fingerprint, face, and iris recognition are not readily deployable in remote authentication schemes. Keystroke dynamics provide the ability to combine the ease of use of username / password schemes with the increased trustworthiness associated with biometrics. Our research extends previous work on keystroke dynamics by incorporating shift-key patterns. The system is capable of operating at various points on a traditional ROC curve depending on application specific security needs. A 1% False Accept Rate is attainable at a 14% False RejectRate. An Equal Error Rate of 5% is suitable for systems requiring a relatively low security. As a username password authentication scheme, our approach decreases the system penetration rate associated with compromised passwords by 95%-99%. Said performance measures can be further improved through optimization of the classification algorithm on a user specific basis.

A Survey on Keystroke Dynamics Biometrics: Approaches, Advances, and Evaluations

“…We discuss recent advances and new trends in keystroke dynamics research. Echoing the sentiment on a lack of common evaluation framework [8,60] in the field, we compile a list of publicly available keystroke datasets. We would like to note that despite the available datasets, we are still in need of large standard keystroke databases for the research community.…”

“…Both classical statistical methods [39,46] and advanced machine learning approaches have been used, including K-Nearest Neighbor (KNN) classifiers [21,112], K-means methods [59], Bayesian classifiers [81], Fuzzy logic [44], Boost learning [8], and Random Forests [8,76], etc. Support vector machine (SVM) is a powerful machine learning method which computes decision boundaries by maximizing the margin in order to reduce the generalization error.…”

Multi-modal decision fusion for continuous authentication

“…The feature space that has been investigated ranges from the simple metrics of key press interval [8] and dwell [9] times to multi-key features such as trigraph duration with an allowance for typing errors [2]. Furthermore, a large amount of classification methods have been studied for mapping these features into authentication decisions.…”

Utilizing linguistically enhanced keystroke dynamics to predict typist cognition and demographics

“…Stylometry describes the measurement of linguistic "style" and has been 12 effectively used in authorship attribution [7,8], in dating a single piece of 13 writing [9] and in establishing genre shifts within the work of a single au- 14 thor [10]. However, whereas keystroke dynamics has been used to verify the 15 identity of one of hundreds of typists, stylometric applications typically dis- 16 tinguish between many fewer individuals. The metrics developed within this 17 field typically rely on the user's spelling of specific words, choice of words in 18 a sentence and choices with respect to grammar.…”