Conceptualization of an Integrated Procedure Model for Business Process Monitoring and Prediction

Becker Jörg, Brunk Jens, Ding Wenying, Niemann Marco

Abstract

Predictive Process Monitoring (PPM) aims to improve operational business processes through the prediction of future behavior and process-related performance indicators. It includes a set of techniques to predict the future behavior of business processes based on knowledge of previously executed process instances. The need for rapid intelligent support in business process management brings growing attention to the field of PPM. A variety of machine learning (ML) based PPM techniques have been proposed by researchers to tackle several PPM-related prediction tasks, such as next activity, time, risk, or outcome. However, adopting such a PPM technique involves many complex tasks, including the selection and parametrization of a machine learning algorithm. Additionally, the novel domain lacks a set of guidelines on how to implement the methods to enable a broad and non-expert user group to benefit from it. As a remedy, this paper proposes a procedure model, which can serve as a step by step framework during the implementation and application of PPM. We conclude our work by demonstrating the applicability of our procedure model and by outlining our case study-based evaluation concept.

Keywords

Data mining; Data models; Task analysis; Adaptation models; Monitoring; Knowledge discovery

Cite as

Details