Abstract
Ontologies are being utilized widely as sources for formally organized information in a range of fields. The SNOMED CT ontology is a key resource in national and international health sectors for automatically linking information captured by diverse clinical information systems and research data ensuring consistent patient data capture and effective data analytics and decision support. Offering a comprehensive multilingual vocabulary for encoding clinical knowledge of multiple domains, the ontology is large and new releases are created regularly to reflect domain changes and user requirements. The main contribution of the paper is a novel automated approach for tracking semantic differences of subdomains in different versions of SNOMED CT targeted at terminologists, debuggers, ontology evaluators and developers of software using SNOMED CT. Whereas the semantic difference sets produced with existing methods are rather large and difficult to analyze, our method produces concise semantic difference sets for user-specified input focus concepts. Our method is based on subontology generation and semantic difference computation using uniform interpolation, which aids in finding inferred differences that other semantic difference tools do not reveal. The obtained semantic difference sets are related to the meaning of focus concept definitions for specific ontology subdomains, where some of these differences would not have been generated without this focused method for computing semantic differences between ontologies. A case study using SNOMED CT has shown the proposed approach is useful for domain experts.
Original language | English |
---|---|
Pages | 250-258 |
Number of pages | 9 |
DOIs | |
Publication status | Published - May 2023 |
Keywords
- Forgetting/Uniform Interpolation
- Modularisation
- Ontology Engineering
- Ontology Extraction
- SNOMED CT
- Semantic Differences
- Subontologies
- Subontology Extraction