Event logs without explicit case identifiers remain a key challenge in process mining, particularly when working with sensor-rich environments. Construction hoists illustrate this challenge: on active sites, sensor streams capture overlapping events, such as floor calls, door cycles, and lift movements, without specifying which ride each event belongs to. The effect of alternative surrogate case-identification rules on both performance metrics and process model structures, especially in noisy, real-world sensor data, remains underexplored in the literature. This paper evaluates six rule-based surrogate identifiers using a month-long dataset from a European hoist rental firm, comprising nearly three million events from 180 sensors. For each identifier, we (i) generate a BPMN model, (ii) compute key performance indicators (KPIs) such as hoist utilization and waiting time, and (iii) measure structural similarity between models. Results show that identifiers can yield comparable KPIs, yet may produce structurally distinct models, revealing variations in captured process behavior. Our contributions are: (1) a reusable catalogue of surrogate case-identification rules, (2) a replicable evaluation workflow, and (3) practical guidance for selecting case-identification strategies in construction and logistics contexts without native identifiers. The choice of strategy depends on the analytical objective, data characteristics, and the balance between interpretability and model detail.
Building similarity graph...
Analyzing shared references across papers
Loading...
Erjan Steenbergen
Rob Bemthuis
Faiza Allah Bukhsh
Procedia Computer Science
University of Twente
Building similarity graph...
Analyzing shared references across papers
Loading...
Steenbergen et al. (Thu,) studied this question.
www.synapsesocial.com/papers/69c37bd4b34aaaeb1a67ea38 — DOI: https://doi.org/10.1016/j.procs.2026.02.155
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: