Business & Information Systems Engineering

Document Type

Research Paper


Artifact-centric process models aim to describecomplex processes as a collection of interacting artifacts.Recent development in process mining allow for the dis-covery of such models. However, the focus is often on therepresentation of the individual artifacts rather than theirinteractions. Based on event data, composite state machi-nes representing artifact-centric processes can be discov-ered automatically. Moreover, the study provides ways ofvisualising and quantifying interactions among differentartifacts. For example, strongly correlated behaviours indifferent artifacts can be highlighted. Interesting correla-tions can be subsequently analysed to identify potentialcauses of process performance issues. The study provides astrategy to explore the interactions and performance dif-ferences in this context. The approach has been fullyimplemented as a ProM plug-in; the CSM Miner providesan interactive artifact-centric process discovery toolfocussing on interactions. The approach has been evaluatedusing real life data, to show that the guided exploration ofartifact interactions can successfully identify process per-formance issues.