Graph embeddings for Process Data

Graph embeddings turn relationships in process data into numerical representations that machine-learning models can use. These works study how to include process knowledge in those representations and apply them to predictive process monitoring.

Enhancing Remaining Time Prediction in Business Processes through Graph Embedding

HICSS 2025.

We use graph embeddings to represent relationships between process activities for remaining-time prediction. These representations can be used with different prediction algorithms and are evaluated on real event logs.

Paper

Vector Representation for Business Process: Graph Embedding for Domain Knowledge Integration

ICMLA 2023.

We investigate graph embeddings as a way to represent business-process data and incorporate domain knowledge.

Paper

All research topics