Related Experiment Videos
Beyond synonymy: exploiting the UMLS semantics in mapping vocabularies
O Bodenreider1, S J Nelson, W T Hole
1Management Systems Designers, Vienna, Virginia, USA.
Proceedings. AMIA Symposium
|February 3, 1999
Summary
This study introduces a method using semantic relationships within the Unified Medical Language System (UMLS) to map concepts between vocabularies, achieving 50-65% mapping success to MeSH.
Area of Science:
- Medical Informatics
- Knowledge Representation
- Computational Linguistics
Background:
- The Unified Medical Language System (UMLS) organizes biomedical concepts using semantic relationships.
- Semantic locality is key to understanding and locating concepts within the UMLS.
Purpose of the Study:
- To develop and evaluate a method for mapping concepts between different UMLS vocabularies.
- To assess the effectiveness of using semantic relationships for concept mapping to the Medical Subject Headings (MeSH) vocabulary.
Main Methods:
- Utilized semantic relationships between concepts within the UMLS.
- Applied the method to map non-MeSH concepts to MeSH.
- Conducted a manual review to assess the relevance of the mappings.
Main Results:
- Successfully mapped 50% to 65% of non-MeSH concepts to MeSH.
- Achieved a 61% relevance rate in the manual review of mapped concepts.
- Identified limitations including inconsistent relationship representation and categorization inconsistencies in the UMLS.
Conclusions:
- Semantic relationships offer a viable approach for cross-vocabulary concept mapping within the UMLS.
- The method demonstrates potential but is constrained by data quality and consistency issues in the UMLS.
- Further adaptations are possible for broader applications in biomedical knowledge organization.