Table 3.

Performance on relation types and overall. Bold indicates micro-averaged F-measures of SR classifier that are significantly different from the corresponding micro-averaged F-measure of the tokens-in-concepts baseline. Italic indicates micro-averaged F-measures of the SR classifier that are significantly different from the corresponding micro-averaged F-measure of the tokens-in-sentence baseline.

BIDMC corpus
Relation type	Tokens-in-concepts baseline		Tokens-in-sentence baseline		SR classifier
	Micro-avgd F-measure	Macro- avgd F- measure	Micro-avgd F-measure	Macro- avgd F- measure	Micro-avgd F-measure	Macro- avgd F- measure
Present disease–treatment	0.64	0.56	0.66	0.53	0.81	0.69
Possible disease–treatment	0.67	0.35	0.75	0.41	0.74	0.40
Present symptom–treatment	0.61	0.45	0.69	0.53	0.80	0.62
Possible symptom–treatment	0.76	0.30	0.89	0.41	0.95	0.47
Disease–test	0.81	0.67	0.73	0.68	0.88	0.80
Disease–symptom	0.93	0.71	0.94	0.77	0.96	0.84
Overall performance	0.75	0.49	0.75	0.54	0.86	0.63
Partners corpus
Relation type	Tokens-in-concepts baseline		Tokens-in-sentence baseline		SR classifier
	Micro-avgd F-measure	Macro- avgd F- measure	Micro-avgd F-measure	Macro- avgd F- measure	Micro- avgd F- measure	Macro- avgd F- measure
Present disease–treatment	0.69	0.42	0.75	0.55	0.78	0.58
Possible disease–treatment	0.63	0.60	0.66	0.52	0.76	0.65
Present symptom–treatment	0.58	0.46	0.62	0.48	0.68	0.57
Possible symptom–treatment	0.78	0.41	0.80	0.47	0.91	0.65
Disease–test	0.70	0.60	0.72	0.65	0.78	0.72
Disease–symptom	0.61	0.56	0.69	0.65	0.69	0.66
Overall performance	0.67	0.49	0.71	0.54	0.76	0.62