Skip to main content

Table 1 Performance of ChER under the BioCreative IV CHEMDNER track setting.

From: Optimising chemical named entity recognition with pre-processing analytics, knowledge-rich features and heuristics

Custom Features

Post-processing

CEM

CDI

 

Abbr.

Comp.

P

R

F1

P

R

F1

✓

✗

✗

92.76

81.02

86.49

91.39

85.29

88.23

✓

✓

✗

92.76

81.30

86.65

91.37

85.45

88.31

✓

✗

✓

92.14

81.41

86.44

90.55

85.72

88.07

✓

✓

✓

92.14

81.69

86.60

90.53

85.88

88.14

  1. Key: Abbr. = Abbreviation recognition, Comp. = Chemical composition-based token relabelling