Automatic inference of indexing rules for MEDLINE
2008

Automatic Indexing Rules for MEDLINE

Sample size: 200000 publication Evidence: moderate

Author Information

Author(s): Névéol Aurélie, Shooshan Sonya E, Claveau Vincent

Primary Institution: National Library of Medicine

Hypothesis

Can Inductive Logic Programming (ILP) be used to infer indexing rules for automatic indexing recommendations in MEDLINE?

Conclusion

The ILP-based approach outperforms manual rules and improves the performance of the Medical Text Indexer (MTI).

Supporting Evidence

  • ILP rules showed higher precision and recall compared to baseline methods.
  • The study utilized a large training corpus of 100,000 citations.
  • ILP rules were assessed against a gold standard of MEDLINE indexing.

Takeaway

This study shows how computers can help automatically label medical articles with the right keywords, making it easier to find them later.

Methodology

Inductive Logic Programming (ILP) was used to infer indexing rules from a training corpus of MEDLINE citations.

Potential Biases

Potential bias due to the training corpus used for rule inference.

Limitations

The performance of ILP rules in a real production environment may differ from theoretical assessments.

Participant Demographics

Citations randomly chosen from MEDLINE.

Statistical Information

P-Value

p = 0.001

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1186/1471-2105-9-S11-S11

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication