Process Area Extraction by Multilevel Resource Detection for Object-Centric Process Mining
摘要
Organizations manage a variety of processes at multiple levels, often simultaneously, creating complex interactions among diverse object types. This multi-level process structure challenges traditional object-centric process mining techniques due to overlapping resource utilization across different process areas. To address this complexity, this paper introduces a novel approach for multilevel resource detection in object-centric process mining, aimed at structuring and clarifying the interactions across processes that may operate on different levels of the organization. We propose a method to detect and categorize process areas based on an automatically detected resource hierarchy, using object types to detect process boundaries effectively. Our approach provides three main contributions: a theoretical framework for identifying process areas by resource levels, a publicly available implementation for practical application, and a qualitative and quantitative evaluation. The evaluations demonstrate the method’s capability to enhance the clarity and interpretability of mined process models, ensuring that higher-level processes are not cluttered by unrelated lower-level activities. This method enables the detection of process areas and resource hierarchies that motivate future research of resource analysis methods that utilize these hierarchies.