Skip to main content

Table 7 Evaluation of AML’s matching pipeline in the Anatomy and FMA-NCI small tasks with different combinations of background knowledge information source (lexical vs. cross-references) and usage strategies (mediator vs. lexical expansion)

From: Tackling the challenges of matching biomedical ontologies

Task BK info BK usage Precision Recall F-measure
Anatomy Lexical Mediator 94.8% 91.0% 92.9%
  (1365-1352) Expansion 94.6% 90.2% 92.4%
  Cross-refs Mediator 93.4% 92.5% 93.0%
  (1389-1401) Expansion 95.0% 93.6% 94.3%
FMA-NCI small Lexical Mediator 93.9% 91.7% 92.8%
  (1624-1598) Expansion 93.6% 91.5% 92.5%
  Cross-refs Mediator 95.8% 91.0% 93.3%
  (1541-1559) Expansion 94.2% 92.0% 93.1%
  1. The number of classes of the two input ontologies covered by each information source is shown within parentheses below the source in each dataset