Selecting event monitoring points for optimal prediction quality

Abstract

Organizations strive to optimize their business processes in order to satisfy customer requirements and internal goals. A basic necessity in order to meet time and quality objectives is to monitor an organization's business processes. Process monitoring makes their execution more transparent and allows to react to observed deviations with corrective actions. This paper focuses on monitoring processes in manual or semi-automatic environments, where the installation of each monitoring point is costly, as it requires effort to measure and record observed progress. During process execution, the allocation of event monitoring points (EMPs) is restricted to certain positions, e.g., the termination of activities. We propose an approach for optimizing the allocation model of EMPs in order to improve the estimation quality. We implemented this approach and show its applicability in a case study of a Dutch hospital for its surgical care process.