Often the observed behavior of a business process differs from the normative behavior captured in its corresponding model. Existing differences can arise from workarounds devised by the workers or regulatory changes not transferred to the model. Process mining and, in particular, conformance checking aims to analyze the observed behavior captured in an event log and to find the differences to the model. Existing conformance checking techniques are the foundation for process model repair — the differences detected by a conformance checker are reconciled in the model, i.e. the repaired model better fits the log.
Through this tutorial, we look into existing process model repair techniques. We focus solely on control-flow information, and broadly categorize the repair techniques into two groups: automatic, and interactive and incremental. While automatic approaches do not require any input from the user, interactive and incremental approaches are used as support tools for the users, who decide what repairs to apply. This tutorial discusses the strengths and weaknesses of both approaches. Furthermore, this tutorial encompasses a brief overview on the conformance checking techniques used as the foundations for the presented process model repair techniques. In particular, it focuses on two conformance checkers, one based on trace alignments and one based on behavioral alignments. However, only a short introduction about the trace alignment-based conformance checker is presented in this tutorial since the tutorial “Conformance Checking: What does your process do when you are not watching?” is fully devoted to this topic.
When and where
Thursday 14th, 16h — 17h (Sala d’Actes, Vertex)
Abel Armas Cervantes is a postdoctoral research fellow within the BPM research group at the Queensland University of Technology, Brisbane, Australia. He obtained his PhD in Computer Science and MSc in software engineering within the Software Engineering group at the University of Tartu, Tartu, Estonia. His research interests include models for concurrent systems, analysis techniques for business process models, process mining, and social network analysis.