This research paper presents a systematic approach to discover true execution dependencies in business processes. Current process discovery algorithms often rely on time precedence, which can sometimes be misleading from a causal perspective. To overcome this, the authors propose a method that leverages an existing causal discovery algorithm to reveal the true causal dependencies among tasks in a process. The methodology is tested using two open process mining algorithms, the IBM Process Mining tool, and the LiNGAM causal discovery technique, on a synthesized dataset and two open benchmark datasets.
Publication date: 23 Oct 2023
Project Page: arXiv:2310.14975v1 [cs.AI] Paper: https://arxiv.org/pdf/2310.14975
