MARC Record Matching with Bibliographic Metadata
Traditional matching of MARC (Machine-Readable Cataloging) records has relied heavily on identifiers like OCLC numbers, ISBNs, LCCNs, etc. assigned by catalogers. However, this approach struggles with records having incorrect identifiers or lacking them altogether. This model has been developed to match MARC records based solely on their bibliographic metadata (title, author, publisher, etc.), enabling successful matches even when identifiers are missing or inaccurate.
Check out the code and dataset at our
GitHub repository.
Try out our
interactive demo to see the model in action!
Key Features
- Bibliographic Metadata Matching: Performs matching based solely on bibliographic data, eliminating the need for identifiers.
- Text Field Flexibility: Accommodates minor variations in bibliographic metadata fields for accurate matching.
- Adjustable Matching Threshold: Allows tuning the balance between false positives and false negatives based on specific use cases.
Performance
This model achieves 98.46% accuracy on our validation set (see our
GitHub repository for datasets).
It has also had comparable accuracy with SCSB, and Goldrush on a separate set of English monographs. Each matching algorithm was run on a common set to produce a union set of all of the algorithms' matches. Using a matching threshold of 0.99 to minimize false positives, we were able to compare the algorithms' matches. Disagreements between the algorithms were manually reviewed, resulting in false positives and false negatives for those disagreements:
| Algorithm | % False Positives | % False Negatives |
|---|
| Goldrush | 0.30% | 4.79% |
| SCSB | 0.52% | 0.40% |
| Our Model | 0.23% | 1.95% |
