Translating message sequence charts to other process languages using process mining

Message Sequence Charts (MSCs) are often used by software analysts when discussing the behavior of a system with different stakeholders. Often such discussions lead to more complete behavioral models in the form of, e.g., Event-driven Process Chains (EPCs), Unified Modeling Language (UML), activity diagrams, Business Process Modeling Notation (BPMN) models, Petri nets, etc. Process mining on the other hand, deals with the problem of constructing complete behavioral models by analyzing event logs of information systems. In contrast to existing process mining techniques, where logs are assumed to only contain implicit information, the approach presented in this paper combines the explicit knowledge captured in individual MSCs and the techniques and tools available in the process mining domain. This combination allows us to discover high-quality process models. To constructively add to the existing work on process mining, our approach has been implemented in the process mining framework ProM (www.processmining.org)