Skip to main content

Table 3 Comparison of BioNER for chemicals

From: BioByGANS: biomedical named entity recognition by fusing contextual and syntactic features through graph attention network in node classification framework

Method\Dataset

BC4CHEMD

BC5CDR-chemical

p

r

f1

p

r

f1

tmChem [25]

89.09

85.75

87.39

–

–

–

TaggerOne [27]

–

–

–

94.20

88.80

91.40

BiLSTM-CRF [28]

91.31

87.73

89.48

92.82

88.52

90.62

Att-BiLSTM-CRF [20]

92.29

90.01

91.14

93.49

91.68

92.57

CollaboNet [29]

–

–

–

94.26

92.38

93.31

MTM-CW [31]

91.30

87.53

89.37

–

–

–

BioKMNER [32]

–

–

–

–

–

94.00

BioBERT-MRC [33]

93.89

91.96

92.92

94.37

94.00

94.19

MTL-LS [34]

–

–

92.42

–

–

93.83

BioELECTRA[35]

–

–

–

–

–

93.60

BioBERT [11]

92.80

91.92

92.36

93.68

93.26

93.47

Proposed

93.42

92.52

92.97

94.53

94.95

94.74

  1. Bold indicates the best performances of models in each subtask