Table 2. Correlation between Mapping Gain and F-measure, between Similarity Score and F-measure, and between Effectiveness and F-measure.

Evaluation Task	Correlation with F-measure
	Mapping Gain		Similarity Score		Effectiveness
	All	no UMLS	All	no UMLS	All	no UMLS
Mouse-Human	0.998	1.000	0.830	0.884	0.628	0.688
FMA-NCI small	0.997	0.988	0.168	0.716	0.965	0.783
FMA-NCI whole	0.609	−0.985	0.800	−0.938	0.613	−0.840
FMA-SNOMED small	1.000	0.994	0.135	0.925	0.997	0.901
FMA-SNOMED whole	0.996	0.944	0.993	0.789	0.994	0.860
SNOMED-NCI small	0.999	0.987	0.682	0.928	0.994	0.591
SNOMED-NCI whole	0.999	0.972	0.982	0.867	0.993	0.581
Average	0.998	0.981	0.632	0.852	0.929	0.734

Correlation coefficients were computed with all background knowledge sources and with all sources except UMLS. The average was computed excluding the FMA-NCI whole task, as the reference alignment for this task is incomplete, resulting in the negative correlation coefficients observed without UMLS.