Skip to main content

Table 5 Comparison between AML’s full matching pipeline with and without the use of Lexicon weights to score the mappings

From: Tackling the challenges of matching biomedical ontologies

Task Lexicon weights No Lexicon weights
  Precision Recall F-measure Precision Recall F-measure
Anatomy 95.0% 93.6% 94.3% 81.3% 95.8% 88.0%
FMA-NCI small 95.8% 91.0% 93.3% 94.6% 91.1% 92.8%
FMA-SNOMED small 92.3% 76.2% 83.5% 90.9% 77.3% 83.5%
SNOMED-NCI small 91.4% 73.6% 81.6% 89.9% 74.1% 81.2%
FMA-NCI whole 83.8% 87.2% 85.5% 78.3% 86.4% 82.2%
FMA-SNOMED whole 88.0% 69.0% 77.4% 84.4% 70.3% 76.7%
SNOMED-NCI whole 89.7% 67.1% 76.8% 86.4% 67.1% 75.6%