Self supervised method improves business process event prediction accuracy
Process-Aware Event Log Augmentation for Process Prediction under Labeled Data Scarcity
Machine Learning
Summary
Predicting the next steps in a business process can be hard when there's not enough example data available. The authors introduce a new method named SiamSA-PPM that creates realistic variations of process event records using knowledge about how processes work. This method learns useful patterns without needing labeled examples, helping it make better predictions about future events and final results in real-world scenarios. Tests show it matches or beats current top approaches in predicting what happens next in ongoing business tasks.
What this means in practice
- •For business process analysts: Improve forecasting of next events and outcomes in business workflows despite limited labeled historical data.
- •For workflow automation engineers: Develop more reliable automated decision support tools that predict process events using augmented event logs learned without manual labeling.$Commercial implications: Enables creation of predictive analytics products for enterprises seeking better process automation and monitoring solutions.
Authors
Sjoerd van Straten, Alessandro Padella, Marwan Hassani
Abstract
Predictive Process Monitoring (PPM) forecasts future events of ongoing business processes based on event logs. However, deep learning PPM approaches are often limited by labeled data scarcity and the low variability of real-world logs. To address this, we introduce SiamSA-PPM, a self-supervised framework combining Siamese learning with statistically grounded data augmentation. It employs novel transformation methods that leverage control-flow semantics and frequent behavioral patterns to generate realistic, semantically valid trace variants, learning generalizable prefix representations without labeled supervision. Extensive experiments on real-world event logs with a wide range of variability, incorporating exhaustive hyperparameter tuning, demonstrate the value of this process-aware augmentation. Compared to state-of-the-art baselines, SiamSA-PPM achieves competitive or superior predictive performance in both next activity and final outcome prediction tasks. An open source implementation of our method is publicly available under https://github.com/SvStraten/SiamSA-PPM-Extension..