Checking Conformance between Colored Petri Nets and Event Logs
Event logs of information systems consist of recorded traces, describing executed activities and involved resources (e.g., users, data objects). Conformance checking is a family of process mining techniques that leverage such logs to detect whether observed traces deviate w.r.t some specification model (e.g., a Petri net). In this paper, we present a conformance checking method using colored Petri nets (CPNs) and event logs. CPN models allow not only to specify a causal ordering between system activities, but also they allow to describe how resources must be
processed upon activity executions. By replaying each trace of an event log on top of a CPN, we present how this method detects: (1) control-flow deviations due to unavailable resources, (2) rule violations, and (3) differences between modeled and real produced resources. We illustrate in
detail our method using the study case of trading systems, where orders from traders must be correctly processed by a platform. We describe experimental evaluations of our method to showcase its practical value.