Foundations of Process Discovery

Abstract Process discovery is probably the most interesting, but also challenging, process mining task. The goal to take an event log containing example behaviors and create a model that adequately describes underlying process. This chapter introduces baseline approach used in commercial tools. A simplified so-called Directly-Follows Graph (DFG). explain challenges one faces when trying discover model. After introducing DFG discovery, we focus on techniques are able models allowing for concurrency (e.g., Petri nets, trees, BPMN models). distinguishes two types of approaches such models: (1) bottom-up (2) top-down . Alpha algorithm presented as technique. has many limitations, nicely idea discovering local constraints. basic inductive approach, combined with frequency-based filtering, works well logs. These algorithms illustrate foundations discovery.