Graph neural networks predict multiple future process events accurately
GNN4PPM: Multi-Target Predictive Process Monitoring with Relational Graph Convolutional Networks
Machine Learning
Summary
Predictive Process Monitoring helps predict what will happen next in a business process based on past events. The authors present GNN4PPM, a new method that uses graph neural networks to look at the relationships between all the data recorded during a process, not just the activities and timestamps. This approach can predict multiple future events and their details at once, improving prediction in complex real-life cases. They tested their method against other top techniques and found it performed well.
What this means in practice
- •For business process analysts: Improve accuracy in forecasting all upcoming events and their details during process execution for better decision-making.
- •For enterprise it teams: Integrate enhanced predictive tools that leverage full event log data to optimize workflow management and resource allocation.
Authors
Ana Costa, Johannes Mäkelburg, Luise Pufahl
Abstract
Predictive Process Monitoring (PPM) aims at predicting at runtime and as early as possible the future states of a process execution. Common tasks include predicting the next event, the time to completion of a trace, and outcomes. Existing approaches typically consider an event from the perspective of the executed activities along with their timestamps and case identifiers. This leads to the disadvantage that in real-life settings, there is much more information recorded in the event log that is not captured or completely ignored when performing prediction tasks. We introduce GNN4PPM, an approach that predicts all next events along with their complete data payload at once. We represent event information in a heterogeneous knowledge graph that captures the event log as an RDF semantics, and train the embeddings with a Relational Graph Convolutional Network (R-GCN). Our approach is promising in comparison to existing solutions, and experiments with state-of-the-art solutions prove the accuracy and applicability of GNN4PPM in complex settings.