1 Introduction
Event sequence data, i.e., multiple series of timestamped or ordered events, is increasingly common in a wide range of domains. Website click streams, user interaction logs in software applications, electronic health records (EHRs) in medical care and vehicle error logs in automotive industry can all be modeled as event sequences. It is crucial to reason about and derive insights from such data for effective decision making in these domains. For example, by analyzing vehicle error logs, typical fault development paths can be identified, which can inform better strategies to prevent the faults from occurring or alert drivers in advance, and therefore improve driver experience and reduce warranty cost. Similarly, by analyzing users' interaction log with software applications, usability issues and user behavior patterns can be identified to inform better designs of the interface.