Predictive Compliance Monitoring Approaches for Violations

This document is a research article that presents two innovative predictive compliance monitoring approaches aimed at quantifying the magnitude of compliance violations in process management. The authors, Qian Chen, Stefanie Rinderle-Ma, and Lijie Wen, identify a gap in existing compliance monitoring methods, which typically provide only a binary yes/no assessment of compliance. The proposed methods reformulate the compliance prediction problem by integrating classification and regression tasks, allowing organizations to not only predict compliance violations but also measure their severity. The article emphasizes the importance of quantifying the extent of deviations from compliance constraints, particularly temporal constraints, which are critical in various application domains such as healthcare and logistics. The evaluation of these approaches on synthetic and real-world event logs demonstrates their effectiveness in providing deeper insights into operational performance, thereby enabling informed decision-making to mitigate risks associated with non-compliance.