Robustness of keystroke dynamics identification algorithms against brain-wave variations associated with emotional variations

Abstract

Keystroke dynamics facilitates the identification of a person by the way they type. This article focuses on analysing the robustness of keystroke dynamics algorithms against variations in biometric records through electroencephalography, using waves associated with states of relaxation and excitement and a self-report questionnaire. An experiment was conducted to capture keystroke patterns in different affective states. The results suggested specific classification distances such as A and R metrics, Canberra distance and two Minkowski-based distances have their accuracy slightly and negatively influenced by changing moods. Euclidean distance seemed to be the least affected.