Abstract
In this paper, we introduce a new approach to recognising typing behaviour (biometrics) from an arbitrary text in heterogeneous environments using the context of time series analytics. Our proposed method differs from previous work directed at understanding typing behaviour, which was founded on the idea of usage a feature vector representation to construct user profiles. We represent keystroke features as sequencing discrete points of events that allow dynamically detection of suspicious behaviour over the temporal domain. The significance of the approach is in the context of typing authentication within open session environments, for example, identifying users in online assessments and examinations used in eLearning environments and MOOCs, which are becoming increasingly popular. The proposed representation outperforms the established feature vector approaches with a recorded accuracy of 98 %, compared to 83 %; a significant result that clearly indicates the advantage offered by the proposed time series representation.