Knowledge Base automatically extracted from songfacts.com following the methodology described in the following paper.

KB Entities Triples Relation Patterns Cluster Patterns
KBSF-ft 20,744 32,055 20,438 14,481
KBSF-th 10,977 11,720 2,484 828
KBSF-co 30,671 113,561 --- ---
KBSF-raw 29,280 71,517 47,089 32,712
KBSF-rv 9,255 7,532 2,830 ---

The dataset includes the 5 KBs generated plus the evaluation subsets used in the manual evaluation (top 100 and random 100). The best version of the dataset (the one with higher precision) is KBSF-th. Further information about the dataset and the evaluation results can be found in the paper.

Download KBSF dataset
 

Scientific References

Please cite the following paper if using KBSF.
Oramas, S., Espinosa-Anke L., Sordo M., Saggion H., & Serra X. (2016).  Information Extraction for Knowledge Base Construction in the Music Domain. Data & Knowledge Engineering.

License

Dataset compiled by Sergio Oramas and Luis Espinosa-Anke. This dataset is licensed under Creative Commons CC BY 3.0.

Feedback

Problems, positive feedback, negative feedback... it is all welcome! Please help me improve KBSF by sending your feedback to: [email protected] In case of a problem report please include as many details as possible.