GNN4PPM: Multi-Target Predictive Process Monitoring Data

Ana Costa's dataset supports the publication 'GNN4PPM: Multi-Target Predictive Process Monitoring with Relational Graph Convolutional Networks'. The pipeline combines RDF graph embeddings with relational graph neural networks to predict process attributes. It is designed for multi-target predictive process monitoring tasks, such as next event prediction.

Use Cases

• Next event prediction in business processes using graph neural networks.
• Multi-target attribute prediction based on relational graph structures.
• Benchmarking relational graph convolutional network models for process mining.
• Analyzing process behavior through heterogeneous graph representations.

Strengths

• Dataset is associated with a specific, peer-reviewed publication (GNN4PPM), providing academic context.
• Released under a permissive CC-BY-4.0 license, allowing for reuse and modification.
• Cross-platform presence on figshare indicates established sharing and potential for community use.

Limitations

• Column names and row counts are not provided in the metadata, limiting understanding of the data structure.
• The dataset size is listed as 2494204184 (likely bytes), but the specific composition of files within the ZIP is unknown.
• No information is provided on the source or generation method of the underlying event log data.

Provenance

Source

Ana Costa

Collection Method

Created to support the GNN4PPM publication; method involves combining RDF graph embeddings with relational graph neural networks.

Freshness

Last updated on 2026-03-20.