Detecting Anomalous Events in Object-centric Business Processes via Graph Neural Networks

By Alessandro Niro, Michael Werner

Rating

1340
Battle Count: 33

Relevance

6/10
While not directly applicable to trading, the anomaly detection approach could be adapted for financial process monitoring and fraud detection

Implementation Complexity

7/10
Requires understanding of graph neural networks and object-centric process mining concepts

Reproducibility

4/5
Code and datasets are publicly available, experimental setup is well-described

About this paper

Methodology: Graph Convolutional Autoencoder. Problem types: Anomaly Detection, Graph Learning.

The interactive Everscope explorer (charts, battles, favorites) loads below.