Skip to main content

Table 9 Recall/precision break-even point (BEP) results

From: Recognizing speculative language in biomedical research articles: a linguistically motivated perspective

Method

Recall/Precision BEP

baseline1

0.60

baseline2

0.76

Our system on the fruit-fly dataset with SA weighting

0.85

Our system on the fruit-fly dataset with IG weighting

0.80

Our system on the BMC dataset with SA weighting

0.82

Our system on the BMC dataset with IG weighting

0.70