Fig. 5
From: Identification of conclusive association entities in biomedical articles
Analysis of the frequency-based indicator (TF): Most entities have TF less than 6 (see the two dashed lines). Most non-CAEs have a TF value equal to 1, and probability gain of finding CAEs (ProbGain) becomes large when TF ≥ 4 (see the solid line). TF thus performs well in identifying CAEs whose TF = 1 or TF ≥ 4. However, many entities have TF values falling between 2 and 3, and for these entities TF has difficulty in distinguishing CAEs from non-CAEs