KGI-Bench: New Benchmark for Knowledge Graph Integration Pipelines
A new benchmark called KGI-Bench has been proposed to evaluate data integration pipelines for knowledge graphs. It assesses pipelines based on coverage, correctness, and consistency of the updated KG. Benchmark datasets for the movie domain include a seed KG, overlapping input data in three formats, and a reference KG as ground truth. The approach was demonstrated by evaluating 12 pipelines across different input data.
Key facts
- KGI-Bench evaluates integration pipelines for knowledge graphs.
- Three quality metrics: coverage, correctness, consistency.
- Benchmark datasets cover the movie domain.
- Includes seed KG, input data in three formats, reference KG.
- 12 pipelines were comparatively evaluated.
- Published on arXiv with ID 2605.22304.
Entities
Institutions
- arXiv