Transfer Learning for Predictive Process Monitoring

Document Type
Conference Object
Issue Date
Liessmann, Annina
Wang, Weixin
Weinzierl, Sven
Zilker, Sandra
Matzner, Martin

Event log data reflects the behavior of business processes that are mapped in organizational information systems. Predictive process monitoring (PPM) transforms this data into value; it creates process-related predictions to provide insights required for proactive interventions at process runtime. Existing PPM techniques require sufficient amounts of event data or other relevant resources that might not be readily available, preventing some organizations from taking advantage of PPM. In this paper, we present a transfer-learning-based technique for PM, allowing organizations without suitable data or other relevant resources to implement PPM for effective decision support. We instantiate our artifact, and apply and evaluate it on a real-life use case. The use case includes two event logs for purchase-to-pay processes of two organizations. Our results provide evidence that knowledge of a process of one organization can be transferred to a similar process of another organization to enable PPM in the target organization.