Abstract
SNOMED CT is a compositional terminology. Construction of post-coordinated expressions allows users to specify new meaning by referencing existing SNOMED CT concepts. The use of post-coordinated expressions in information systems requires special software, a reasoner, to give the exact relations between post-coordinated expressions and existing SNOMED CT content. Thus, the performance characteristics of reasoners are important for implementation of post-coordination in information systems. This study aims to test how reasoners perform when a large number of post-coordinated expressions are added to SNOMED CT. The time needed to classify an ontology consisting of SNOMED CT plus an increasing number of post-coordinated expressions is measured. The best performing reasoner in this test classifies SNOMED CT plus 1 million post-coordinated expressions in 42 seconds. The time to classify grows a little less than quadratic as the size of the ontology increases. In conclusion, classification time is not a problem using current reasoners and current SNOMED CT releases even if a large number of post-coordinated expressions are added
Original language | English |
---|---|
Pages (from-to) | 1048-1052 |
Journal | Studies in health technology and informatics |
Volume | 205 |
Publication status | Published - 2014 |