Predictive Process Monitoring: a comparison survey between different type of event logs

TL;DR

This paper addresses how Predictive Process Monitoring (PPM) can leverage different event log types, notably object-centric event logs, to forecast future process behavior. It conducts a systematic literature review of 25 studies, classifying approaches by Event Log Type, Prediction Task, and Method to reveal how OCEL-based PPM differs from classical-log PPM. Findings indicate OCEL-based methods are newer and rely on graph representations or feature engineering to capture object interactions, while classical-log methods favor DL/ML models for higher accuracy but with tradeoffs in explainability and training time. The study highlights gaps in OCEL benchmarking and provides a structured foundation for future benchmarks and cross log comparisons to drive practical PPM deployments.

Abstract

The application of Predictive Process Monitoring (PPM) techniques is becoming increasingly widespread due to their capacity to provide organizations with accurate predictions regarding the future behavior of business processes, thereby facilitating more informed decision-making. A plethora of solutions have been proposed in the literature employing these techniques, yet they differ from one another due to a number of factors. However, in light of the growing recognition of the value of object-centric event logs, including in the context of PPM, this survey focuses on the differences among PPM techniques employed with different event logs, namely traditional event logs and object-centric event logs. In addition, the reviewed methods are classified according to the prediction task they address and the specific methodologies they employ.