In user behavior mining (UBM), process mining algorithms are applied to the user interaction logs (UI logs) of an IT system. These are fine-grained records of interactions between users and software, such as clicks or keystrokes. By applying process mining to this data, previously unrecognized patterns in user behaviour can be identified and visualized. Despite the existing similarity between UI logs and process-based event logs, the application of process mining algorithms to UI logs is not easily possible. In the paper, we analyze existing challenges of UBM and define an agenda for future UBM research.
This paper is the result of our ongoing cooperation with SAP Signavio. The authors are Jana Rehse, Luka Abb, Timotheus Kampik, Christian Warmuth, Carsten Bormann and Gregor Berg. The entire publication can be read here.